26OH-2021-08-15_mixdown

Harpreet: [00:00:05] What's up, everybody, welcome. Welcome to the comet Emelle Machine Learning OpenOffice hours powered by the artist Data Science Superexcited. Have all of you guys here today shout out to everybody in the room asking what's going on? Christoph Parap A..

Harpreet: [00:00:20] Ahmos, if you guys joining

Harpreet: [00:00:22] In on a LinkedIn or if you join again on YouTube, twitch wherever you may be joining in from, feel free to go ahead and let us know if you have any questions. We'll be taking questions for the next hour or so, or maybe even longer, depending on how fun the conversation gets. Hopefully I've got a chance to tune into the episode, or at least on the science podcast earlier this weekend. On Friday, actually, I spoke to Jaclyn Wale's. We talked about her book, The Fearless Factor. I felt like it's a pretty personal conversation. I think I was coming to a lot of stuff that I was dealing with at that time. I interviewed her back in probably February and during that time and I was going through some severe, like kind of burnout issues. I was going through some stuff and and she was there just talking, talking me through it. So it was a really, really good episode. I enjoyed speaking to her and enjoyed getting some of her tips. We touched on a lot of different topics that we talked about, you know, struggling to follow through, struggling with Forell follow through and consistency. How do you overcome that? We talked about systems and why systems are better than goals. Talk about self awareness. Um, you know, we talked about why should stay open and curious, not judgment about four different characteristics of self-awareness as well. Um, we can to ourselves are good enough. It get upset, man. Like I really enjoyed speaking with her to hopefully get a chance to check that out. Also shout out to uh to Matt rattin, Matt Bratten if you guys see them on, if you guys are joining in on LinkedIn tag. Matt, Brian right there on the comment, he sent me this awesome shirt. Uh uh. So for those who are listening on the podcast, it's they select from where shirt. Uh, so he sent this to me, um, [00:02:00] that I don't know. Hey, man, appreciate you saying that to me. Uh, Matt, really, really, really, really enjoyed this shirt.

Harpreet: [00:02:06] Um, and so

Harpreet: [00:02:07] We'll go ahead of starting questions. You guys have questions. Go ahead. Put them right there into the chat. Um, go ahead and put them into the comment section. Wherever it is at watching from. I'm keeping a good eye on, uh, on all that stuff. Um, shout out

Harpreet: [00:02:20] To, uh, to Junior,

Harpreet: [00:02:21] Who's joining us from Brazil on LinkedIn. Happy to have you here. So let's go out and get warm up with a question. Actually, I've got a question that, uh, that came

Harpreet: [00:02:28] In through, um,

Harpreet: [00:02:30] Through email. And it's a question from, um, one of our community members who,

Harpreet: [00:02:35] Uh, uh, whose name is three

Harpreet: [00:02:37] Six six two

Harpreet: [00:02:38] Collective, uh, and

Harpreet: [00:02:40] I don't see an actual name on the email, but they're saying, uh, I see you get a lot of advice from all the people you interview in your podcast. Do you have a framework for taking in and implementing the advice you got? I thought that was a really good question, and I wish I had a really good

Harpreet: [00:02:55] Answer for that. But I don't

Harpreet: [00:02:57] I don't really know that. I don't think that I have a framework

Harpreet: [00:03:00] For from implementing

Harpreet: [00:03:01] The advice that I get. Unfortunately, what I try to do is try to just internalize it, let it sink in and connect to other ideas I have in my head. Um, and that's not the most efficient way. Like, honestly, like I've been on this kind of journey of this, uh, personal growth and development and just trying to become better only the last three years or so. And my process up until then has just been

Harpreet: [00:03:23] Literally highlighting

Harpreet: [00:03:25] Pages in my book. It's been lagging my book with that, with notes or just writing in the book. And it's gotten to a point where I've reached a critical mass of books like I've had a ton of

Harpreet: [00:03:36] Books which with a ton

Harpreet: [00:03:38] Of flags in them, but none of them are organized, are systematized, or I have no real way to really, interestingly connect ideas to, uh, to each other. Um, it's all having to happen in my

Harpreet: [00:03:49] Brain, which isn't the best way to do it.

Harpreet: [00:03:52] So I've been reading this book recently, just finished it. It's called How to Take Smart Notes. Um, I think I might have been talking about this earlier before, um, and, [00:04:00] uh, probably about a month or so. I've been I've been doing this for a while and just really digesting it. And, uh, what this book is based on, it's this method that's called a total kastin method. And it's by, um, it's invented by a professor, I think of it is in Germany. Uh, last name I think is looming. And he is a professor of sociology. And it's been really, really helpful for me recently. And I'm going to be going over the course of the next several months, going back over through some of the books that I've read and taking those notes from the highlighted sections with the flag sections and really implementing them in that whole cast. And so what that is, is, you know, I'll start by uh, I recently started by taking the leading notes and literature notes, and then from there, I'll move them onto a

Harpreet: [00:04:49] Little, uh, and

Harpreet: [00:04:50] Notecard like this. I've got a five by notecard and then I start writing on five. I know card and I put it into place right here. Which holds all the five I know cards, and once it's in there, my plan is to go through at least once

Harpreet: [00:05:07] A week and put them

Harpreet: [00:05:08] Into a personal knowledge management system. And I've been exploring a few different options for that. So one of them was Roehm Research. One of them was a notion. The thing about the man, like they're super expensive. But I recently came across something called Obsidian, and I'm not sure if you guys are familiar with Obsidian or if you've heard of Obsidian, but it really helps facilitate this personal knowledge management system. You're able to add tags to your notes. You're able to add bidirectional links to your notes. And it's really cool because inside the app itself, it you can see like a knowledge graph of all the different types of things that you study and learn about and all of the text that you write, all the notes that you write, they're just simple markdown files. And so quite it's quite cool to use.

Harpreet: [00:05:58] Interesting, uh, idea. [00:06:00]

Harpreet: [00:06:00] I wish I would have came across this much earlier in life, but then again, much earlier in life was kind of kind of maybe not the case anymore, but yeah, definitely. Check that out, Obsidian. There's a couple of huge resources that that I recommend for learning

Harpreet: [00:06:16] About is that

Harpreet: [00:06:17] Kastin learning about how to use software like Obsidian after Zettl Castle. There's just one Zettl Kastin Easy says he leakiest he and Deasey. And then those guys have like an entire long form blog just breaking down how to use the system again. There's also the note, the book, How to Take Smart Notes that'll teach IDL the telecast method. But for using Obsidian specifically to resource that I came across in the last couple of weeks, one has been, um, linking your thinking on YouTube. So linking your thinking on YouTube has this really comprehensive, like from the ground up tutorial on how to use, um, Obsidian and then Brian Jenckes on YouTube who were recently connected with on on LinkedIn as well. It turns out that we're from the same part of Sacramento as well. I thought that's pretty cool. Uh, so he's got Brian doing some YouTube's got a whole mess of stuff on productivity and using Obsidian for productivity. Um, so if you link obsidian up with another app called Soltero, you're able to really quickly and effectively start creating references and bibliographies and stuff like that. So I recommend checking those things out. Um, I, um, I'm brand

Harpreet: [00:07:36] New to it, like brandnew like, you

Harpreet: [00:07:38] Know, two, three weeks into it. But I can already tell it's going to be a

Harpreet: [00:07:41] Game changer for me.

Harpreet: [00:07:43] And I'm really, really excited about going through some of these books. Now I've got hundreds and hundreds of books, well, maybe just one or two hundred books laying around with just notes and flags and them that I need to get out and into a system. Uh, so, yeah, that's that's enough warming up there. So I'm, uh, I'm keeping an eye out on all the [00:08:00] chats and all the comments section. If you guys have questions or comments, please do put them right there into the comments section. I am more than happy to take your questions, guys. Um, but yeah. So hopefully you guys get a chance to look into, uh, some of that software. And also, hopefully you guys are following along with twenty one days of deep learning that's been really fun and really cool to do. So, Betty, how's it going? Betty, I see your unmuted. So if I got a question, go ahead. Let me know. All right. Betty does not have a question tour. How's it going? Pretty good. Yeah. I can hear you

Tore: [00:08:32] As usual, sitting outside a local office

Harpreet: [00:08:36] And sitting outside in the in the south of France, uh, suffering with a beer and have

Tore: [00:08:42] A little, quote unquote, heat, you know, and that's bit warm and cool.

Harpreet: [00:08:52] I see a question actually coming into the chat right here, actually in, uh, in Zoom's. So let's go ahead and get to that. Um, it

Harpreet: [00:08:58] Is from pirate,

Harpreet: [00:09:00] Uh, saying any specific or recommendation or someone transitioning from a software engineer background to a machine learning engineer background as a follow up, what would be some great projects, particularly in NLP health care, which showcased to boost our chances of applying for a normal engineer role? It's a couple of different parts. Maybe. I think that question first thing, I think, um, uh, there's of course, that Mikiko, one of my friends, Mikiko, is recommends and that's called a full stack, deep learning. And I can go ahead and I'll pull that up right here, actually,

Harpreet: [00:09:33] On, uh, on

Harpreet: [00:09:35] It I on the screen here. But we'll stack deep learning. It's, um, it's completely free and it's, uh, taught by, uh, some answers at, uh, I think UC Berkeley. It is, um, and it's all about

Harpreet: [00:09:51] The

Harpreet: [00:09:51] Engineering and deployment of machine learning projects.

Harpreet: [00:09:54] Uh, so this

Harpreet: [00:09:56] I think be the probably be best stauss. For somebody who's coming [00:10:00] from a,

Harpreet: [00:10:01] Um, software engineering

Harpreet: [00:10:02] Background, trying to get into machine learning, engineering, this seems like a really, really good, um, program for that. So I highly recommend for

Harpreet: [00:10:08] Stac Deep Learning

Harpreet: [00:10:10] And Mikiko herself has transitioned from a Data scientist role to a proper machine learning engineer role, and she credits a lot of her learning and

Harpreet: [00:10:19] Success to that

Harpreet: [00:10:20] Transition to this most active learning course. So highly

Harpreet: [00:10:24] Recommend that now in terms of a

Harpreet: [00:10:26] Great project that could showcase your chances for playing, I mean, great projects are always going to be the ones that you are absolutely most interested in. Right. So those are the best projects. I don't think there's like any one particular project that you could do that might. Wow.

Harpreet: [00:10:42] A hiring manager, right?

Harpreet: [00:10:44] What what will our hiring manager is just your execution of the project. How do you write your code? How do you do your write ups? How do you take somebody from point A to point B? Um, how well documented is your code? Right. How well thought out is the work that you're doing? Um, so that's how I would answer that question. Like, I don't know if there's a great project speaking of NLP or health care, like, um, you want to do an NLP project and I was supposed to work on this Cristoff. Maybe we could, uh, touch base on this later, but we're thinking about taking all the transcripts from all the happy hours and doing something with that using NLP. Because we've got I mean, I've got over one hundred transcripts from almost a hundred transcripts from like fifty something happy hour sessions and like almost thirty comment more sessions.

Harpreet: [00:11:30] So that's a lot of text Data

Harpreet: [00:11:32] There that you can use to do something with. I'm sure. Um, in terms of health care, I don't really have,

Harpreet: [00:11:38] Um, much

Harpreet: [00:11:39] Knowledge about the health care domain. I mean, I used to work in, um, biostatistics, the pharmaceutical industry, which I think is a little bit different from health care. Um, but Cristoff, what do you think what are some interesting projects that you've been working on for for NLP?

Cristoff: [00:11:52] I wish I knew the answers, but I didn't I didn't actually work on NLP so deep yet. [00:12:00] I used some like I mentioned, I used some libraries like spacy to to know a lot about linguistics or to extract some information from text Data when I used it was so my project idea was to get articles from NBA games because that's what I'm interested, interested in NBA basketball. So I just scraped, uh, like recaps and did some projects. I'm from the Data on the text Data. But like Harpreet Sahota it must be interesting for you.

Harpreet: [00:12:40] Yeah. And I mean, one thing you might want to do, like if you're trying to get into machine learning, engineering is I mean, definitely experiment management is going to be easy for that.

Harpreet: [00:12:49] So check out some of the projects that, um, comment

Harpreet: [00:12:52] Has up on their website. I could pull that up or ask them to drop a link to that. That'll be great. Um, they've got we've got a ton of projects that not in the near future will be coming out with a lot more as well to kind of, uh, teach you guys a little bit more on experiment management and module production monitoring, things like that. Um, all right. Are you still here? Was that was that helpful? Do you have any follow up questions that, uh, yes.

Baharat: [00:13:17] Harpreet Sahota is really helpful. Thank you. And, uh, yeah. One more follow up question. Since you mentioned that you may be considering pulling up the transcript from the happy hour sign, probably the officer.

Harpreet: [00:13:33] So if you could

Baharat: [00:13:35] Make that public or even semipublic the way that we can access it, then yeah, that would be really helpful because that's another real area of interest. But I have hit a snag, like

Harpreet: [00:13:48] Even

Baharat: [00:13:48] From other content creators. I tried putting, like, uh, transcripts from Spotify or LinkedIn, general transcripts from podcasts I was trying to take from you for [00:14:00] your podcast.

Harpreet: [00:14:01] They should be there on my podcast. We go to just three hours a day of science, go to any individual episode. You have to go to the website itself. If you go to the episode itself, um, and like I'll show you real quick if you go to the

Harpreet: [00:14:13] Episode like this. Right.

Harpreet: [00:14:15] So this is like the most recent happy hour that was set up. And just go to a transcript and it'll pull up a text file. That's the entire transcript. Oh, OK. We'll have that. I have that for every single, uh, every single one. So there's a lot of manual effort. Um, yeah. I mean, I actually don't have any the transcript stashed on my on my hard drive. I have it just set up so I can quickly.

Baharat: [00:14:36] Oh. Mean, this is fine. We can probably automate this process using some scraping tool. I'm not sure if scraping is allowed, but using some manner of some tool which does something similar to scraping. So even this is fine. I didn't even see this option on

Harpreet: [00:14:55] Some people's part.

Baharat: [00:14:56] Karslake Only the watch. The initial intro was president

Harpreet: [00:15:03] And other

Baharat: [00:15:03] Creators podcast, so that's why they had trouble, but thanks, I did

Harpreet: [00:15:08] That itself would be a cool, like full on N10 project where you're sourcing Data. So story going to place your source in the Data and then you're saving that somewhere, maybe not no single database or something like that, and then doing transformation's to get it ready to be modeled and stuff. And then finally Buruma like that's complete end to end. That is that would be a good, good project man.

Harpreet: [00:15:29] Yeah. Thank you. That that looks very

Baharat: [00:15:32] Like, you know, a promising thing. At least I find it interesting because that's like little raw data out there. And you can do whatever you want with it. It's going to be very nice.

Harpreet: [00:15:43] Yeah, yeah, yeah. I've been I've been

Harpreet: [00:15:46] Pausing on on that NLP project just because I've been doing this twenty one days of deep learning thing. But, um, I know there's so much knowledge that could be done. I think with those text transcripts that I have, um it'd be interesting to see, uh,

Harpreet: [00:15:59] What happens [00:16:00] if you do have some

Baharat: [00:16:02] You do plan to do some sort of collaboration and definitely open with respect to these transcripts, at least because, you know, I been wanting to do this kind of project.

Harpreet: [00:16:12] Yeah, yeah, maybe we could think of something that could, uh, we could do and do some experiments on a comment like that, I think that

Mark: [00:16:17] I was going to say I was going to say Harp, this sounds like a first maybe the seeds of a first project for your time at Comet. Sounds interesting to me.

Harpreet: [00:16:25] Yeah, definitely, man. I think it'll be a lot of fun, uh, to Mark Marks in the building. How's it going, man? Good for you, Mark. Actually, I was asking about questions. I started a question around a project idea for health care. I know you've got some health care background. I was wondering if you might have like a, uh, interesting project idea. I know you're talking about something that, uh, you guys did, um, on on on Friday. So you bring that up again.

Mark: [00:16:52] Definitely. So I think one one quick question is, are you trying to do a data analysis project or are you trying to do a machine learning project? So is this a question of LinkedIn or someone in the group?

Harpreet: [00:17:04] This is a group from Borith right here.

Mark: [00:17:06] Oh, hey, it was a hey.

Baharat: [00:17:09] So it's on the second part, the machine learning part and learning.

Mark: [00:17:13] Ok, so the machine learning part, it's a little bit harder. And I think the reason being is that like you, often they take a step back. There's there's a lot of health care, there's a lot of health care. And so for those like I think the thing you probably want to, like, really look into is like what area and health care do you want to look at impact on? And the type of Data are you looking for

Harpreet: [00:17:36] Like tabular

Mark: [00:17:37] Data or are you looking for like more unstructured data, like patient notes or computer vision? This once before I go down the rabbit hole of like advice.

Baharat: [00:17:47] Ok, then then in that case it would narrow it down to patient, not because they want to

Harpreet: [00:17:55] Do it and do

Baharat: [00:17:56] It. NLP.

Mark: [00:17:57] Ok, so yeah. NLP project [00:18:00] around patient notes. OK, so I think the challenge I imagine you've probably already tried looking for Data and in health care and it's like extremely sparse. And a big reason for that is that you have a patient kind of patient protection, which is a great thing. Right. So what you could potentially look for is like moch datasets around patient notes. So biosynthetic Data I'm trying to find real world Data health care is extremely hard. And to be honest, I would, just because of the security implications, actually feel much better working with synthetic Data. So if I was in your shoes, I would look up at a synthetic patient note Data Data said, if that's possible. And so one of the key things like the pain problem point is that health care Data is extremely messy and so they are Data AIs to work with you. Notes do a lot of text data mining and that and the challenge is that like the way that electronic health records are captured, it's captured by disparate levels of systems. So you have like one hospital using this system for electronic health records, another hospital using another system, another hospital, another system, and then they come together. And it's just a complete mess. In addition, you have different doctors describing things in different ways. And then furthermore, on top of that is that sometimes doctors like I think one of the problems you have is like find all the surgeries that were conducted in this time frame, one of these patients. And then you'll have like I conducted I conducted the surgery or I mentioned the surgery or

Harpreet: [00:19:44] I

Mark: [00:19:45] The patient declined the surgery. Right. And so that's the kind of problem and use cases like making sense of Data of like really, really messy and kind of unstructured and standardized [00:20:00] Data. So that's a really interesting NLP project. And so diving for adventure, that's that's the pain point. Any questions before I dove in a little bit further?

Baharat: [00:20:10] No, at this point, I still understand it because I

Harpreet: [00:20:13] Did do an

Baharat: [00:20:14] Internship already, what did do some sort of extraction. It was information expection exactly like this, even that we worked on the same synthetic patient lots. But yeah, I use a little different. But it did involve going through, like, all of the patient notes and then actually getting things very patient. So, yeah, I totally understand what they're saying and this is what they want to dove deeper into.

Harpreet: [00:20:41] Amazing, amazing.

Mark: [00:20:42] So cool. You already have the context of how challenging it is. And so one of the key things that why you can't just throw a random NLP model, you download spacy or not, you can just throw it at it, is that it has really contacts and domain specific knowledge. And so I know Spark NLP as a health care package. And so I will look for specifically like health care specific NLP packages that start off on. And then more importantly, it's like don't try to boil the ocean, focus on one specialty. So where it's like ophthalmology and neurology. And then within that you saw the data set available, focus on one disease that will scope it down to where there's already a lot of variability in that. But then you can just focus on that and I'll get you a solid start for India. And I think that's where most people get tripped up on working on health care. Data, especially around NLP, is that they try to boil the ocean. There's so much variability, so much different domain knowledge.

Harpreet: [00:21:39] If you scope

Mark: [00:21:40] It down to a specific problem, use case will be set up for success. Hopefully that was helpful. I know that just like a giant dump of health care information.

Harpreet: [00:21:48] Oh, thank you.

Baharat: [00:21:49] That really helps. You know, like I mean, like just meeting clues and searching for resources. And they know like I'm facing the same difficulties because [00:22:00] they're confidential for the most part and protected, if not confidential. So exactly. This is what I was looking for. Thank you. Thanks bye.

Mark: [00:22:09] I think I can remember that our notes are not, but I think they have I think the CDC look up the the synthetic entrepreneur health care data set, I think is presented by the CDC. But I use it whenever I work on the last couple of times. I work on trying to build my own startup around health care. You need Data to build a proof of concept. And we use those data sets are massive. They're like they're like millions and millions of records. You cannot run it on on your local computer, but you can download it, get a sample of it and start playing around with it. I believe they have data set in there and it's all synthetic

Harpreet: [00:22:43] And it's another resource here as well, the simple math, Data said. Austin posted. Here's a link in the mass dataset that's available for download as a IT as a million synthetic patient records. So I think between those two resources, there should be enough to to get something started, something fun going on.

Mark: [00:23:04] It looks like the one that I share in the two. There's some specific use cases around, I think, childhood diabetes and orthopedics oncology as well. So there's a couple of different specific covid-19 as well. So there's some specialized subset are like data sets within that

Harpreet: [00:23:17] Synthetic

Mark: [00:23:18] Sort of patient records and health care stuff there.

Harpreet: [00:23:22] Yeah, it's pretty cool. Um, well,

Baharat: [00:23:25] Thank you all. Thanks a lot. I mean, this is like a lot of to take didn't even know existed.

Mark: [00:23:31] So if this person is a child. So that wasn't the CDC. Is this CMS

Harpreet: [00:23:38] Center

Mark: [00:23:38] For Medicaid and Medicare Services? I believe so. That's where they have that that data sets synthetic data set as well.

Baharat: [00:23:45] Ok, yeah, I just copied the link.

Harpreet: [00:23:49] So this is this is great. And I didn't know that there is all this wonderful type of health care data out there. I was I just assumed it was private and locked away and that we couldn't [00:24:00] get anything close to the real world. So.

Baharat: [00:24:04] Exactly. So, like,

Mark: [00:24:06] As a quid pro tip, a great data structure to work with, this massive data sets are files. That's what I typically use. I'll create a quick function. I'll convert Iris Kesby into a file and condenses it down to like ten megabytes compared to like one hundred or something like that. Look at PRK files are here. You eat. It's a great tool working with when working on larger data and you're trying to hack something together. Work locally.

Harpreet: [00:24:33] Yeah, good files. Better than

Harpreet: [00:24:35] Pikul. Another good one is better, but I don't think feathers are used that that frequently in most people's is right up. And that should be a good, good starting point for there man. Excellent question. Yeah. Yeah.

Baharat: [00:24:50] Excellent resources actually. Thank you. I love you. Like I didn't even know these datasets existed frankly.

Harpreet: [00:24:57] This is. Yeah.

Harpreet: [00:24:59] Awesome. Well if anybody else has questions, go ahead. Let me know. Shout out to everybody else in the room. So ammos or Betty. Two names. I've not seen this in the Harp. Our story officers were happy or rather yet happy to have you guys here a.m., but if you guys have questions, please do let me know. Um, shout out also to a.. And or and Christopher, you guys have questions. Let me know. Keeping an eye out on, uh, on all the comments sections everywhere. I don't see anything coming through. Um, so. So you just kind of just wave and we can take without any without any burning questions, just get them started. So, um, not that popped up in my head while I was doing the happy hour last week. We're having some discussions and something pop up to my head. And I wrote this down on the side here. And I think I'd love to get some perspective on this, but let's say there's these legacy companies that are going through digital transformations. Right? Should a data scientist be a part of a digital [00:26:00] transformation? And if so, what role does the data scientist play? Right. And again, this is coming from just my experience working at Legacy Company and trying to be a part of a digital transformation. And it's tough to do. And I don't know if if that

Harpreet: [00:26:16] Was the right

Harpreet: [00:26:17] Place to to put a data scientist into somebody who's kind of spearheading initiative on digital transformations and things like that. Just something I'm putting out there. See if we can get some discussion going on around that. But what

Harpreet: [00:26:32] Do you think Mark

Harpreet: [00:26:33] Marks the most thinking it through? I don't know, man. For me personally, like, I feel like, yeah, to a certain extent they should be a part of a digital transformation,

Harpreet: [00:26:41] But,

Harpreet: [00:26:42] Um, maybe not leading the entire charge on one, because that is a lot of work for a skill set that's probably not best suited for doing something like that.

Mark: [00:26:52] So I recently read an amazing article. I'm gonna go try to find input in the chat, but they discuss the same exact problem where they made up a fake company and a fictional story around driving Data transformation. And so essentially the core of the story was that they have this Data team that we're just building notebooks and it wasn't going into production all there. It is kind of like a cost center of typical thing where the new companies like we need Data science and they hire a whole bunch and then nothing happens. And so they hire this leader

Harpreet: [00:27:24] Where they

Mark: [00:27:25] Actually start implementing throughout the culture, like driving Data maturity and driving that Data infrastructure. And a of is centered around SQL and game people used to using SQL and understanding Data governance and whatnot. And one of the key problems was that for his Data science team, they essentially were like, well, I'm not doing Data science anymore, so I'm just going to leave. So you had people

Harpreet: [00:27:46] Leave, you had people who are

Mark: [00:27:47] Data analysts who aren't jumping the Data science rolls. Right. And so it really came to this decision like we need to build Data infrastructure. We have data scientist to help, but they're not get the more experience I can promise [00:28:00] that will come to them in like a year. But right now it's just not possible. Right. And so that was kind of the core of the story is like driving Data infrastructure. You're going to have to do the work ahead of time before you start doing the fun stuff. But it's important that you bring in someone with that skill set. So that way you can help guide that people to that direction. And so as a data scientist, I read that story. I was like, wow, I've been on all of those shoes. Like I've been I've been the person trying to drive the changes in the company. I've been the data scientist. I'm like, oh, man, not really doing data science. Right. But now I've also been a data scientist like really committed to the mission. And I want to see it through because that's just an awesome project to say. Like, Hey, I built this Data infrastructure. I helped drive this culture and now I'm doing data science I think I built. And so bringing it back from nonfictional to like my own story is that I'm kind of in that role right now at home where I get to do some really cool Data science things. But allow my work at a startup is like building the infrastructure to allow us to do those more advanced things. I could push in now, but that would just set me up for failure because I'll create a model and we'll go into production then. That was Komaki waste a lot of time money on that. And so for me, why it is interesting for me is that I want to build startups, I want to build my own company. And so I'm sticking around because I love to learn. What do I need to

Harpreet: [00:29:24] Do to build

Mark: [00:29:25] A Data focused startup? What are the steps I need? Right. So my career goals go beyond just being a data scientist. Is this data scientist and leadership and building capacity. And so I'm getting the skill set that I couldn't get out of like a more advanced company. Right. And so that's where I currently stand on it. I think it really comes down to your career goals. And I don't think the the company is not going to be aware of this, especially if you're the first Data science person being hired.

Harpreet: [00:29:52] They're not going to know how to communicate this.

Mark: [00:29:54] So this is going to be on you to basically piece out the information, pull the information yourself. [00:30:00] And so being a first data scientist, I don't think that's a really ideal move for a newcomer because you're not going to get the Data skills you need. But if you're more experienced, then you're going to know, like I have these Data skills, I can work on the side. But this is the strategy I'm implementing. I think able do. That's a real career jump start if you want to go to more leadership roles and not just be. I see. Yeah.

Harpreet: [00:30:26] Yeah, I like that. I love that perspective. And yeah, it's the first. Scientists know it's super hard, I would not recommend that unless you've got some significant experience in a rule like that, you learn a lot, you grow a lot because there's a lot of challenges, but you expect it to do pretty much everything, which is

Baharat: [00:30:45] Tough not to get lost.

Harpreet: [00:30:50] Yeah, I was gonna make sure you're on mute if you're not speaking

Harpreet: [00:30:54] That we don't pick up any

Harpreet: [00:30:56] Background noise. I'll give you a link to that article. Do I love to read that?

Harpreet: [00:30:59] Ask them. Go for it.

Mark: [00:31:00] Yeah, I have thought about this. It might be a little

Harpreet: [00:31:02] Bit

Mark: [00:31:04] Of a sideways entry into this, but I was thinking about someone I know kind of just like keep it more anonymous here, but like someone who is working at a health care company and sort of they were it's sort of like this cart before the horse thing. I think you were getting at Mach where

Harpreet: [00:31:18] They wanted to build like a sort of a

Mark: [00:31:20] Data science data analytics team, but with the express purpose of our company around the express purpose of turning research into a marketing material and insights. And I think that's sort of like this

Harpreet: [00:31:35] Leadership

Mark: [00:31:36] Leadership of a company wanting to hire a Data team without really doing that infrastructure work at first, so that there's no understanding of where how you're going to access the data or where it's stored or who has access to water, why? And especially at a health care company. So I kind of thought and that was that I think there can be a tendency, especially a bigger legacy, companies maybe to sort of think

Harpreet: [00:31:57] Of it as input

Mark: [00:31:59] Output, kind of like [00:32:00] more traditional software engineering, like we present this problem, you engineer a solution and you get this output versus sort of like a building the infrastructure first so that you can understand what Data you have, what it could possibly be used for, and then creating a strategy based off of that. I don't know if that's exactly what you were talking about, Mark, but it's sort of made me think of that where I think a lot of legacy companies want to they know they have all this data and it's like, what do we do with it? We want to turn it into something and go like kind of jump in line and go towards like we want an output from this Data versus like really going from the building blocks of like, how do we understand where our data is coming from, what it looks like, how it needs to be cleaned, who should have access to all these questions? And you just want marketing insights out of it. That's just something I've seen in this particular instance. And I'm imagining plays itself out in different companies as well that just have a lot of money and want to throw a lot of money at a problem like that or just people or whatever it is.

Harpreet: [00:32:55] It sounds very familiar to me that situation. It's tough, though, man. It's tough. Um, so I got some questions trickling in here into the chat. So we'll go ahead and start taking some of those. If anybody has questions, go ahead. Drop me the comment section, wherever it is you're watching, whether that's YouTube or LinkedIn. I got an eye out on on the comment section there or just drop it right here into the chat. I'll be happy to take on your questions. Tor has a question about block chain and document sharing. Um, probably not the best person to ask about that, but I will tell you who is, uh, Carlos Mercado. Uh, so definitely go ahead and and link up with Carlos Mercado, send him a message. He's, uh, the resident Data science block chain expert. He's got some awesome stuff going on in the, uh, block, Genspace as well, and just his overall interest in centralized finance. So I'd probably be looking for that.

Tore: [00:33:52] If you take stock, it's a message. Yeah, definitely. Uh, they don't call it short term memory without a reason.

Harpreet: [00:33:59] Yeah, there's [00:34:00] a there's a link right there from Mark to Carloss login page. So, guys, check that out.

Mark: [00:34:06] I recommend checking out his book as well. You were on the central decentralized finance. It really got me up to speed on the basics of block change. Yeah, yeah. It's a great

Harpreet: [00:34:18] And there's also some good free courses on LinkedIn learning as well. A couple of them by, um, I'm probably gonna mess up his last name, Jonathan, until I actually interviewed him for my podcast as well. They'll be released sometime in the future. But he's got a lot of really, really cool courses on LinkedIn learning, um, that help introduce people to the concepts of blockin and cryptocurrency and all that stuff.

Harpreet: [00:34:44] Um Cristoff has a question Cristoff.

Harpreet: [00:34:45] Go for it. And then by the way, if you guys do have questions please let me know. Keep an eye out and all these chats. Uh. Yes, a good question. Now, how do you avoid burnout? I mean, that's tough, man. I went like I said, I was going through some pretty severe burnout earlier this year. And it because I did not take rest, I did not take breaks. I was just going, going and going.

Harpreet: [00:35:14] So I think that is probably

Harpreet: [00:35:16] The key to avoiding burnout is to take a break and take take a break. If you do feel like you're doing too much, like you got to the point where I was waking up every morning, just my head was hurting. I was getting like six hours of sleep. And I, um, again, cynical about everything. And then at that point I was like, let me take a break. But I didn't know what burnout was. I didn't know what it was until I started reading. Reading a book, Max Renzo's book is a, uh, a researcher. He wrote a book called A Time

Harpreet: [00:35:44] Off Him and

Harpreet: [00:35:46] Dsdj.co/artists Name, but that's what I really saw with him. It's going to be in a couple of weeks all about burnout. Um, and I was reading that book and I was like, holy shit, man, I'm suffering these symptoms, these burnout symptoms, um,

Harpreet: [00:35:59] Because mostly [00:36:00] it

Harpreet: [00:36:00] Didn't have a good

Harpreet: [00:36:01] Rest. Epic betore.

Harpreet: [00:36:03] It sounds like you have some tips for avoiding burnout because it sounds like he's

Tore: [00:36:08] Just he's just like, yeah, it's OK to say no. Because generally speaking, you have to know your own limits and try to learn your limits and signals this stuff out there, that kind of thing. But the key is, you know, people will accept it if you need extra time to get yourself that. That's what I've been through the burnout once in my life, but not doing that so well. I think it's time to take a break, do something else. And like I said, people start pushing, asking. That's one of the expectations that you're trying to deliver on other people's expectations. Not. So that's that's the key.

Harpreet: [00:37:03] Yes, I like that. Good, simple, actionable advice. What about us, the Nordmark native

Harpreet: [00:37:09] Reporting growing up?

Mark: [00:37:10] And it's tough. I think one of the difficulties of this is that burnout has now become sort of a thing that people have accepted that it's OK to talk about, but then inevitable. So that's good. And there's a shared struggle in that, and especially after the last year and a half of it and lockdowns and work from all the stuff. But then I think the flip side of it is that you're then you're also seeing everything pop up around how to fix burnout, or there's all these, like little niche markets that pop up about like self care and wellness. And even the idea of taking care of yourself has become commercialized all of a sudden and has become like fit into the mold of like take care of yourself so you can get back to work. And I think that is such a dangerous way to think about this. So it's a tough problem. I've always found that that idea of self care and wellness is always really [00:38:00] about is really hyper

Harpreet: [00:38:00] Personal, and it's about identifying

Mark: [00:38:02] The things in your life that that sort of make you feel connected again. So you were talking

Harpreet: [00:38:08] About

Mark: [00:38:10] The NBA earlier? I think about little things like when I'm feeling out or this sort of type of thing, I feel less inclined to call people and talk on the phone with them. And like, simple things like that are really valuable to me and they make me feel better. But it's harder to get over that hurdle. So I think it's like a really I think it's hard to sort of have a general sweepings sort of thing about how to avoid burnout. But I think it's really like taking an honest look at yourself and identifying those things that make you feel more connected to family or more connected to your interests and really just like really pushing yourself to prioritize those things, and especially in those moments where you're feeling like the work is or your curiosities dimmed or whatever, or it feels lower or whatever. I think it's a hyper personal thing, but I think steering away from something like that, like the

Harpreet: [00:38:55] Commercial like

Mark: [00:38:57] Suggestion's or here's a product you need like don't add more products into your life to solve burnout. Don't add more apps like just don't do it. Like just find those things that are personal to you that make you feel more connected to your immediate surroundings as well, where you live, the people in your community or groups that you're a part of, whether that's digitally or or physically. That's kind of my suggestion. It's not it's sort of there's a bit of abstraction there and not quite one hundred percent helpful. But I think it's a deeply personal thing that has been tried to make it seem like it's fixable by products. And I just don't think that's that's quite the right answer either.

Harpreet: [00:39:32] Yeah, true men don't get enough burnout. There's no app that can really care that much about you.

Mark: [00:39:40] Oh, man. Burnout.

Harpreet: [00:39:41] It's been like a

Mark: [00:39:43] Theme of my early, early 20s and up until now I'm like now going into my late twenties and I'm just now being like,

Harpreet: [00:39:53] Oh, it's OK not

Mark: [00:39:54] To work all the time. And so I've I've bi pattern all [00:40:00] the way from undergrad, grad school to my first

Harpreet: [00:40:03] Few jobs, grind

Mark: [00:40:05] Eighty hours a week,

Harpreet: [00:40:06] Burn out and then

Mark: [00:40:07] Recover or do bad work for a week or two and then just like 80 hours a week. And I was just in this cycle. Right. The thing that really broke that cycle for me was just like being friends, like therapy, like really working with a therapist to be like, hey, why do I have these patterns? Why do I keep on going through? This is like a is like a self-destructive thing for me. And so what I've learned through all those lessons is why I'm like I love towards advice is to say no to things. So I say no.

Harpreet: [00:40:39] A lot more and

Mark: [00:40:40] More importantly is I find what's my 70 percent capacity and try not to exceed that because you're going to want to fill up like, oh, I have all this extra time, let me do it all. The reason why I say 70 percent is that you always have this black swan thing that pops up in your life that you weren't expecting and that gives you a buffer to be like, OK, I'm stretched thin, I'm not going to burn out. Right. So that takes time to figure out what my 70 percent is. And then in addition, the final thing is like sleep. That's the key thing for me personally avoid. Our now and it's like this, this weird thing where, like I have all these deadlines, I'm just going to sacrifice sleep and stay up late to finish it, and then you stay up late, you finish it, and then the next day, because you're still sleep deprived, you're so unproductive. So you work slower. Then you're like, oh, I didn't get the work done. Let me stay up late. And you start this cycle over and over again until you're, like, completely burnt out. And so for me, when I find myself feeling like, oh, my God, I stay up late, I'm like, do 10:00, go to bed, like, break the cycle.

Mark: [00:41:43] And that's the thing that helps helps me avoid burnout. But with that said, like, you're still going to have hiccups along the way. I literally burned out last week. I was overcommitting myself. I, I, I stuck to my 70 percent and I added 40 percent more. [00:42:00] So I was like at one hundred, ten percent capacity. I said way too much. I said yes to too many things. Right. And I, I so what I did I was like, I was like oh wow. I'm really burnt out, I'm really frustrated. I can't focus at work. And so Friday night through this morning, I just did nothing. I just slept and just read and hang out. I watch Lord of the Rings, my wife, which is like four hours long for one movie, and so really forced myself to take a step back, rest and then write one more more. I write, I get back to work, which is kind of counter to Austin's advice of like self care to work more. But I think for me, I enjoy my work so much that it's really hard to balance that. But I question that sometimes. So I like Austin's advice as well yourself.

Harpreet: [00:42:48] Go for it.

Cristoff: [00:42:49] So first of all, great comments. They're very valuable and I think it's very unfair that you still in your 20s because they're still in your 40s or 50s. And I think you said there are different types of burnout or anyone said there are different types of burnout.

Harpreet: [00:43:16] If there's different different types, I'm sure there are varying degrees of it. I mean, maybe there are different types of burnout, but

Cristoff: [00:43:22] Because right now, I don't know if I had a burnout because I didn't work too much. It was actually the opposite because I felt like I was way below my productivity levels, because I felt my job wasn't fulfilling or anything like that. So I just must start procrastination to extreme levels. And I told myself that I had a burnout. Right now, I'm not sure if it was a burden because I found my job purposeless. It's I just hated it [00:44:00] and I didn't want to do that anymore. And I'm just thinking about it right now. It was somewhat better, although it doesn't much definition.

Harpreet: [00:44:09] Yeah. I mean, it sounds like burnout to me. Burnout combined with job dissatisfaction, combination of those two. It looks like a.. Says there's three types of burnout. There's overload under challenge and neglect. Um, so I did not know that.

Harpreet: [00:44:24] So that's that's good to know man.

Harpreet: [00:44:26] Overload under challenge and be like I think when I was going through some stuff earlier this year, it was definitely just the overload part of it. And when I was just in, I felt like that, that I was just putting too many artificial up pressures on myself. And that was that was what caused me to burnout rate. It's like all man. Like I

Harpreet: [00:44:46] Have to listen to an audio book in the shower

Harpreet: [00:44:49] And on my walk and have to use all my time productively, like, can't take a break. I can't waste time. No, can't do any of that. And that led to me just being super, super weirded out. And now when I, when I find myself like OK, like if I feel like now listening to an audio book and just listen to music, that's cool man. I'm just listened to by nineteen nineties alternative rock music. That's the stuff I like, the stuff that makes me feel connected.

Harpreet: [00:45:11] Um but

Harpreet: [00:45:12] You just got to give yourself a break when you start feeling about that, just thinking that everything at least for me go for it

Mark: [00:45:21] Actually does have a more. So a question for you. I listen to Jonathan Tesser podcast with you earlier today or yesterday. I was really good. And a big part of that was talking about like the self-improvement aspect. And so I guess, like, how is the self-improvement aspect? Tie in with the Bernau and baseliner last hour. It kind of sounds like maybe you took the self-improvement to extremes because the amount of waste like was the lesson from all of that.

Harpreet: [00:45:47] You had to do that for me.

Harpreet: [00:45:49] I was just doing what it really was, that case. I took that self-improvement thing to the extreme and I started taking everything that I read in books as if it was a prescription. Right. Like I was just following the recipe. But really, [00:46:00] what I what I'm starting to realize now is, OK, well, I, I just should take bits and pieces of what resonates for me from these different books that I'm reading and just do something for myself that works right. Like me having to write in like the longest time. Writing like four or five different journals, like, you know, like but like I had to do that, like, that's the thing I'm supposed to do. I need to be doing that. I need to be waking up and

Harpreet: [00:46:21] Following the Robin

Harpreet: [00:46:23] Shamos prescription that he lays out in the five a.m. club. And I'll be doing that. Otherwise, I'm not going to be successful. Right. I was I wasn't doing self-improvement just for the sake of improving myself. I thought if I do these things, then I will become successful. Right. Like, the reason that I'm not satisfied or happy in life is because I don't feel successful. And if I want to become successful, then I need to do these things that all these people are telling me to do. Otherwise it's not going to happen. And that sounds like artificial

Harpreet: [00:46:49] Professionals like myself.

Harpreet: [00:46:51] And if that answers your question

Tore: [00:46:57] Successfully, you're all for other people's expectations of

Harpreet: [00:47:03] My own, because I got super high expectations of myself, especially when, you know, like, look, man, like I wasted a lot of my youth, right. Like, you know, I in my twenties, I grew up mostly like, you know, in Sacramento, Sacramento. I think Mark might be familiar with my neighborhood and what it's like and how easy it is for somebody to just, you know, fall off on the wrong path. And that was definitely me. And I wasted a good decade of my life. And I feel like, OK, now it's time to catch up. Now it's time to get back on track and start correcting. Um, and I felt like the only way for me to do that was just to put all this insane amount of pressure on myself to just quickly, quickly get that I can't make the future arrive

Harpreet: [00:47:45] Any quicker by just

Harpreet: [00:47:47] Spinning my wheels in my head. I mean, like, it was just this weird thing that in such a question,

Tore: [00:47:53] Go with the flow. And that's where it got scales. You had sex [00:48:00] that one year. I would just go with the flow. Yeah, it's cool. You have to have fun. And we had fun that was actually and so forth was very costly actually may lose things because of this particular and that missed the opportunity to that. Yeah. And to me it's like going to the solo I said this to this is not part of it and I follow that. And at the same in the past, I want to sell it because if that's stuck in your mind, everything else fall into place.

Harpreet: [00:48:54] Yeah. Yeah, I agree with that.

Harpreet: [00:48:58] Awesome. Go for

Harpreet: [00:48:58] It. And after Austin will go to uh to Marc.

Harpreet: [00:49:00] I think one thing I've been

Mark: [00:49:01] Saying and as we've been having this conversation is that. There can be a tendency when we talk about these kinds of issues to either to sort of paint them in extremes like the one on one sense, burnout is a completely internal self thing like that you need to just overcome, which seems like a fallacy, a bit of a fallacy to me or on the other end is to say this is 100 percent external into these systems. It's my employer, it's my relationships, it's this, that and the other that's causing the burnout. And I think what's really important is it's always more nuanced than that and it's always somewhere in the middle. And I think as you kind of reflect on those feelings or what's going on, it's the sort of way those things try to find a way to weigh those things, at least in a balanced way, so that you're not putting all of this guilt on yourself and you're not projecting all of it outward, because I think there's a tendency, especially when you're feeling cynical or especially when you're feeling down or burn out or whatever it is to really burrow into one of those options. Mine happens to be that I push things out externally because I don't want to deal with the internal failings of my personality. [00:50:00] But I tend to say, like, oh, it's the system, it's this job, it's the industry, it's this it's and there's truth in that, just like there's truth in that. It's an internal thing as well that I struggle with to figure out my boundaries and define my boundaries. So it's like you're sort of I think I think it's really important when you're to take stock of all of that is to really, really try to balance those things as much as you can to really sort of locate the source of this so that so that you're doing it in a way that is holistic as much as possible. It's tough. It's not easy, but that's that's something I've been thinking.

Mark: [00:50:33] Yeah. I think Hossan just touched on it. And I'll say Harp with your comment kind of remind me of this, just like the role of guilt's driving your behaviors and unhealthy ways. And specifically you're talking about growing up in Sacramento. And like I definitely for a brief moment, got a car around the wrong crowd as well. And so my crazy work ethic happened after that.

Harpreet: [00:50:53] So I was like, wow, my

Mark: [00:50:54] Mom sacrificed everything for me. I almost threw it all away. And I that's horrible. So, like, I'm going to work and crazy hard to make sure it doesn't happen, to make sure I don't throw away my mom sacrifices. Right.

Harpreet: [00:51:07] And and what was

Mark: [00:51:09] Funny is like I actually had a conversation. My mom, like a couple of years ago, I was on my path of like unlearning all these bad habits. And she's like, my happiness is for you to be, like, happy and doing something you love, not working yourself to death. And then we had that conversation. I was like, wow, like I didn't work this hard. I still work hard. But the level at which I wasn't working hard to work hard, I was working hard to avoid that sense of guilt and not just facing that. And that's all I was talking about. The therapy,

Harpreet: [00:51:39] Like having a

Mark: [00:51:40] Trained professional to talk you through that guilt is so helpful

Harpreet: [00:51:45] That that's powerful.

Mark: [00:51:47] I was literally typing out a comment about therapy being part of this equation. Margarite, when you said the word therapy is like literally on a comment about how important that is to sort of internalize these realizations [00:52:00] or really put them in balance because it's so

Harpreet: [00:52:01] Hard to do alone, because

Mark: [00:52:03] There's the limits to how your mind and your body can process these things. And having these trusted sort of sources who don't necessarily know the inner workings or your past, your childhood or this or that can be super, super important to help kind of figure out some of these balance points in New Orleans. It's a

Harpreet: [00:52:19] Great discussion

Harpreet: [00:52:21] That kicked off Cristoff. Thank you very much.

Harpreet: [00:52:23] Um, yeah. Yeah, man

Harpreet: [00:52:25] Burned out the burn. That's the thing with aggression. You need to and don't be so hard on yourself. Like says take it easy. Like I was taking stuff to the extreme. I was taking this philosophy stuff to the extreme.

Harpreet: [00:52:37] Uh, I

Harpreet: [00:52:39] Being just like the be the epitome of a stoic and that was just doing it the wrong way and putting more some other philosophies into my life and finding some combination of those to help me have an operating system to deal with this crazy thing we call reality. That's where I was helpful. Anyways, I got a question

Harpreet: [00:52:58] From every

Harpreet: [00:52:59] Night. Are you still here? Yes, you are hurt. You want to go ahead and go with a question or do you want me to read it out?

Harinath: [00:53:06] Yeah, it would be fine if you made it out.

Harpreet: [00:53:08] Ok, great. Yeah. So how are you not the fan that has just recently graduated and gotten a job as a software engineer, but during college was focused most of the time on deep learning, machine learning, uh, finding it really hard to focus on, uh, his job whenever you see stuff that's related to Data trying to change a career path. So I

Harpreet: [00:53:29] Suggested a best

Harpreet: [00:53:30] Way to approach recruiters. Um, I mean, if you you've got if you're working as a software engineer but focused on deep learning, machine learning in college, I think you're already kind of set up for success. Right? I don't know if you call in based out of India. I got to say, look, man, like

Harpreet: [00:53:47] I don't know if the advice I'm giving is

Harpreet: [00:53:49] Going to be applicable to India. It's an entirely different country, different culture value system. Right. So you can take what I'm saying with a grain of salt, like I'm focused mostly on North America and stuff [00:54:00] like that. But the way I would suggest approaching a recruiter is.

Harpreet: [00:54:04] If you see a

Harpreet: [00:54:04] Posting for a job that you like, first of all, just go and apply for it and then try to find somebody in the company, whether it's the technical recruiter or hiring manager, just to say, hey, look, you guys have a job that's posted up, went ahead and applied for already. Let me tell you a few reasons why I feel like I can add a tremendous amount of value to the company. Right.

Harpreet: [00:54:25] So just start

Harpreet: [00:54:26] Applying for jobs and started just talking to recruiters. Right. Nobody's going to I don't think anybody's going be like, oh, well, you're a software engineer. What do you want to do with learning people doing stuff like that? That's not that those aren't incompatible things. I mean,

Harpreet: [00:54:37] So you're going to have to

Harpreet: [00:54:39] Go through the interview process just like anyone else. Right? So you going to prove that you know how to do the stuff as you go through that interview process.

Harpreet: [00:54:45] So I think the

Harpreet: [00:54:46] The way to approach recruiters is just start applying for jobs. And once you apply for a job, go to the company website on LinkedIn. Just send out a message like, hey, look, you know, this is why I think I'm a good fit. Go for it.

Mark: [00:54:57] Yeah. So that's why I'm iterate like, this is not my advice. I'm just repeating something that Mikiko said in other other office hours, but it really resonated with me. So she's an engineer now. She used to be a data scientist and she had this killer resume that was really good. But she wasn't getting Emelle engineer interviews at all. So she talked to a few. My engineer friends like you have an awesome resume for a data scientist, not an engineer. And so when you're applying, they're thinking of you as a data scientist, not not the role. And so she said when she finally switched your resume to make it come off as a male engineer and it essentially like open up the floodgates and she got started all these job offers. So I guess the question I have for you is like, is your resume set up as a software engineer or is your resume set up for, like, the role you want, whether it's male engineer, Data scientist?

Harpreet: [00:55:49] That's an excellent point as well.

Harpreet: [00:55:51] So and I think to

Harpreet: [00:55:53] Make your resume, to be kind of set up like a

Harpreet: [00:55:57] Machine

Harpreet: [00:55:58] Learning engineer, I suppose software engineer is just [00:56:00] focused as much as possible on highlighting the Data, the machine learning related work that you've done in previous roles, whether it's that current company, previous companies or through personal projects and things like that. Right. So I like more of the machine learning aspect of the work that you've done previously going forward. So not let's open it up to you for follow up. So please go ahead and let us know if you have any questions.

Harpreet: [00:56:25] Oh, thank

Harinath: [00:56:26] You very much. First. Actually, Mark, what Mark said was like, that's the thing I'm facing right now. And actually my

Harpreet: [00:56:34] Résumé is in such

Harinath: [00:56:35] A way tied to a deepening project. And I also published a research paper that I think I don't know, like why is and I don't get any back on the accounting stuff like, okay. Something bad or something. So, yeah, that's a problem.

Harpreet: [00:56:51] So make sure you're highlighting those as much as possible and not only on your resume, but when you reach out to a recruiter. Right. So the first thing is if all you're doing is applying for jobs

Harpreet: [00:57:01] And just like playing

Harpreet: [00:57:03] And I hope that my resume gets noticed, like you're just sending it through a digital

Harpreet: [00:57:07] Abyss. You know,

Harpreet: [00:57:08] Who knows if somebody is going to actually look at the resume. Right. But you need to make sure you're reaching out to people really need to be more proactive. Right. That just kind of increases your chances. So let's say,

Harpreet: [00:57:18] For example, you apply for a

Harpreet: [00:57:20] Job that, I don't know,

Harpreet: [00:57:21] Facebook. Right.

Harpreet: [00:57:22] And you submit your resume online, but then you don't do anything, no follow ups or anything like that, then you're not really taking much of that job search process into your control. Right. So you can move the needle just a little bit in your favor by trying to find a technical trying to find the technical recruiter attached to that particular job posting and just shoot them a message and say, hey, look, I've applied for the role already, but let me give you a quick rundown on why I'm the right person for this job, currently working as a software engineer. But I've got a lot of experience in deep learning, machine learning. I've written [00:58:00] a paper that where we discuss this, this and this topic, which is very similar to what it is that you guys are describing in the job posting, going back about a ton of value to a company. Here's my

Harpreet: [00:58:10] Resume. Let me know

Harpreet: [00:58:12] If you'd like to set up a time to call right.

Harpreet: [00:58:15] Or the time to chat on a call.

Harpreet: [00:58:17] So that would be my advice there. Right. So are you doing that? Like what's your process like when you're applying for jobs? Like what do you do? Just applying and that's it.

Harinath: [00:58:27] Yeah. Actually that thing you said are flowing up to the recruiter. I've been actually trying to do that recently. And in that search I found you like when I kinda looked at the data base and I was very thankful to you

Harpreet: [00:58:42] For collecting

Harinath: [00:58:43] Linkedin. So I did. What I usually do is that I applied my me in that company job or whatever, and I tried to apply and like actually I find the job offers directly in the case, but by then finding another not LinkedIn either. So I like I, I don't usually pull this thing like approaching Data. To our message or something, because I was new to this cause they were all and all. I don't know the actual process and I like how to put myself out there and let people recognize me. I have that ability to look for that. I can go.

Harpreet: [00:59:20] Yeah, yeah.

Harpreet: [00:59:20] So definitely if

Harpreet: [00:59:22] You're applying on the actual company website, just go to Data, LinkedIn as well and try to find the same posting on LinkedIn, because typically it'll have like the recruiter that's attached to it right there on the job posting. And then when you do reach out, please don't send a message that just says hi, because if all you do is send a message, says hi, nobody's gonna respond to you. So make sure you just put all the information that you need to convey in one message and make sure that your LinkedIn profile is looking clean. Make sure that you've got you know, you got a professional looking picture, your whatever LinkedIn head header, background picture looks nice and clean professional. [01:00:00] Make sure that your job descriptions mean all your stuff. They're like, it's looking good, right? That's how you can get noticed. Make sure there's a link to your GitHub, link to your research papers. Um, make sure your About Me section is captivating. These are all things that you can control. There's so much in the job search process that you can control to make yourself look like a better candidate. Right. But if you focus on optimizing that, then you don't necessarily have to be at the mercy of what they call the applicant tracking system or anything like that. All of a sudden, you'll notice that opportunities will start coming to you. Right. So if you're

Harpreet: [01:00:38] Linkedin resume looks clean, if you're

Harpreet: [01:00:40] Active for posting insightful comments and your insightful post and, you know, building a brand for yourself, you'll start to get noticed more and more.

Harpreet: [01:00:49] So hopefully that was helpful. Yeah.

Harinath: [01:00:51] Ok, yes.

Harpreet: [01:00:52] Thank you. Thank you. Thank you for doing that.

Mark: [01:00:56] I put a link to a post I made exactly on this topic where I just create a whole template for your email to recruiters. It's my go to template to get responses from from most people.

Harpreet: [01:01:06] Yeah, that's a really good template too. I remember seeing that, um, looking at right now as well. Yeah. I mean that's what it's all about. And like it's completely OK to reach technical recruiters.

Harpreet: [01:01:14] I wouldn't people

Harpreet: [01:01:15] I would not reach out to, I wouldn't reach out to an individual contributor, data scientist

Harpreet: [01:01:20] Just because I don't really have that much clout yet.

Harpreet: [01:01:24] Right. And nobody's really going to be that comfortable with just providing a reference to a recommendation for a complete stranger. Right. If they have to put their own reputation on the line. Um, so the people you should, by all means, reach out to is the recruiters, technical recruiters especially. And then if you see somebody that's got like a leader title attached to their to their name on LinkedIn, you know, manager or lead or director or something like that,

Harpreet: [01:01:53] I'd reach out to them

Harpreet: [01:01:54] Rather than an individual contributor or scientist.

Harinath: [01:01:57] Yeah. OK, thank you. Thank you, Mark, [01:02:00] and thank you. I just wanted to I guess, follow up you guys next week about networking

Harpreet: [01:02:07] And it's also a numbers game. So just make sure you just keep on applying as much as possible. Like, uh, I have got like

Harpreet: [01:02:13] I've got like this this thing

Harpreet: [01:02:15] That I do to help temper my expectations with any job. It is that I apply to it. I just tell myself with any

Harpreet: [01:02:21] Given job application

Harpreet: [01:02:23] Once I submit the resume, there realistically is probably like a one percent chance that if there is a one hundred parallel universes just by me submitting to the resume in one of them, I might get the job. And then as I move through the process, I'll update my probability. Right. Like, OK, I've cleared the ground. Right. There might be a five percent chance that I get this job. I agree. I made it past the first, uh, technical scrape, second technical screen. And up until it gets close and close to the to the end of the process, I just keep updating that probability. But I never really let it exceed a certain threshold. And that's just something I do personally. Uh, it's not like like it sounds Bayesian in the way that I'm approaching it, but there's nothing mathematical about what I'm doing. There's nothing statistically rigorous. It's just the way I manage my own expectations of life. I guess it's the way to put it right. Mark says I'm going to retire. Yeah. So I got this this thinking framework from the book thinking. And that's really helped me that any book that helped me really understand how I can use probability in my own life. And then a couple that with this, uh, research I've done about parallel universes and multiverses and things like that, I'm like, oh, man, doesn't it make sense? I might sound crazy, but at least I'm able to to stay centered and manage my own expectations so that some actually questions coming in from LinkedIn. Finally, uh, there's a comment coming here from Muhammad saying nice hair style. That's Mohammed, like your hairstyle. Um, Jiah Jaya is asking if we can point her to a GitHub resource that uses NLP [01:04:00] to analyze Yelp reviews. Um. I don't have one right off the top of my head, but let me just kind of see if we can find one together. So this is how I'd go about doing it, right.

Harpreet: [01:04:11] The guys being able to

Harpreet: [01:04:13] Search on Google is there is a

Harpreet: [01:04:15] Skill set that I think

Harpreet: [01:04:17] Every data scientist should have. So I got to do is go Yelp reviews and then we could just say file type that we're looking for. I Pinebrook. I Pinebrook. Yes. And then there's a bunch of stuff that will come up and you can just look through them and see which one is helpful.

Harpreet: [01:04:34] Like, OK, like

Harpreet: [01:04:35] For example, if this was some this project that they had submitted as part of the portfolio, I probably wouldn't bring this person on an interview just because I'm pretty well done.

Harpreet: [01:04:45] So either there's like no

Harpreet: [01:04:46] Clarity as to what they're thinking. Um, but if all you're looking at is just some ways to help yourself understand how to approach problems, that's probably how I would go about doing that. So just Yelp reviews, you can even put them at feeling English processing and see what comes up. Um, but I don't I don't want to have one off the top of my head that I can help you up, but maybe we could chat online and find some for you. Do you guys have

Harpreet: [01:05:11] Any, uh, Mark's

Harpreet: [01:05:12] Got a good link here about this.

Mark: [01:05:14] Yeah. So it's shameless plug. I have a tutorial. I did a whole NLP analysis on my own LinkedIn Data. So it's not Yelp Data, but you can easily take all the lessons

Harpreet: [01:05:24] From this and

Mark: [01:05:26] Apply it to apply it to Yelp. And so I walk you through how to do a whole

Harpreet: [01:05:33] Essentially

Mark: [01:05:34] Bag of words, NLP

Harpreet: [01:05:35] Analysis using

Mark: [01:05:38] Using LinkedIn Data. But you can easily switch it out for free.

Harpreet: [01:05:42] Yelp and this is a example of a project that will definitely get you noticed because look, look how clean that really file looks so nice. It tells me everything I need to know. And a quick rundown before I even get to, uh, to the notebook. Then once I get to the notebook, I'm like, oh, yes, he's telling me [01:06:00] what he's thinking. And it's just so much more enjoyable to read through.

Harpreet: [01:06:05] Um, so those are some great resources.

Harpreet: [01:06:08] Also, shout out to Eric Sims in the house.

Cristoff: [01:06:10] Eric, how's it going go? I was on LinkedIn so I can still straighten up and say, hey, thank you.

Harpreet: [01:06:16] Hey, Eric, if you need a good resources for, um, the NLP projects with

Harpreet: [01:06:23] The Yelp review stuff, let

Harpreet: [01:06:25] Us know that Giallo in the LinkedIn comments, uh, help her out. Um, but I hope that's helpful. If not, uh, holler at me on Saturday. We'll see what we could find out for you. Uh, another question coming in here from a cold, uh, chastity, though, is saying he's missing. You guys from LinkedIn here trying to break into Emelle. Any advice on how to learn better and become more employable? Interesting Data. I have no professional background in coding, but I'm currently learning Python and it must be great. That's so funny because

Harpreet: [01:06:55] I literally launching a course

Harpreet: [01:06:57] Called the Employable Data Scientist and another one further down the line about how to learn but how to learn better.

Harpreet: [01:07:03] I've got two immediate

Harpreet: [01:07:05] Resources that you can go to, um, that are going to be shameless plugs for myself. One of them is the interview I did with Scott H. Young. He wrote a book called Ultra Learning to go and listen to that podcast episode with Scott Young, uh, called An Ultra Learner, and also

Harpreet: [01:07:22] Get his book, Ultra Learning.

Harpreet: [01:07:23] It's all about, um, essentially a little faster. Then also Barbarically did an interview with her. She did an entire online class. It's they open online class on Coursera. It's like the most popular or most enrolled in class on Coursera called Learning How to Learn. Those are great resources as well as well as anything by Jim Quick. Jim Quicks, Book Master, um is good as well. So those are great resources. And I didn't give you any advice. I just give you a bunch of resources because, um, there's a big question. Uh, I would say, I would say

Harpreet: [01:07:58] Start by going through

Harpreet: [01:07:59] Some of those [01:08:00] podcast episodes that I mentioned and, uh, put you off and a good start to become more employable. Well, you

Harpreet: [01:08:06] Know, more projects. They do more projects. Right.

Cristoff: [01:08:09] Um, what you guys or something in there?

Harpreet: [01:08:11] Yes. Please go for it. So learn better. It's just like be stupid,

Cristoff: [01:08:15] Like just recognize like you don't know a bunch of stuff and it's you know, like I just like yesterday posted on LinkedIn because I saw from another LinkedIn post something like SQL like window functions. And I was like, I don't even know what that is. So I went, I looked it up and it was like first value of last value. This is going to be super helpful for me and my work. And so I wrote up a query and I shared it on LinkedIn. And I always have this moment of thinking I am the only person that doesn't already know this like this is this is dumb. Why am I even sharing it? But then I go ahead and share it anyway. People are like like, hey, thanks for sharing that. I've never seen it either. It's like, you know, I, you know, I'm learning. And the way that I learn is by sharing and talking at work. Together and just recognizing, like, I don't know, stuff, lots of

Harpreet: [01:09:06] People don't know stuff, it's cool

Cristoff: [01:09:08] To share it and everybody works together and if they already know it, I've never had I've never had anybody on LinkedIn tell me that's stupid. Everybody already knows that, you idiot. Everybody is supportive and helpful or they provide new resources. It's great. So a way to learn better. It's not like learn in public

Harpreet: [01:09:24] For like I've heard that I like to ask people

Harpreet: [01:09:27] Like that learn in public. I think that I've heard Austin talk about that as well. And I mean, that's what I'm doing with deep learning as well. Right. And it's great because you'll get a lot of interesting questions and those questions will make you make you think,

Harpreet: [01:09:39] Um, but, you know,

Harpreet: [01:09:41] To give some quick, actionable advice to, uh, to draw here real quick, I'm going to pull up this carousel that I've made from

Harpreet: [01:09:50] Stuff on

Harpreet: [01:09:51] Jim Quix book I like his framework is called Learning Faster. Right. First thing kind of similar to what Eric was saying is just forget forget what you already know. And [01:10:00] I've got this posted somewhere on my LinkedIn as well. But I can also send you this directly. You'll be happy to. I'll just post it again today. Um, forget a four act as for state, because all learning is state dependent. So once you go into a learning session, boost yourself that you're happy and convince yourself that it's something that you actually want to do that's actually good. Use your time. Don't go into it thinking it's a chore

Harpreet: [01:10:22] Tea, which is similar to what

Harpreet: [01:10:24] Eric is saying, like learning and public try to teach it to someone else. I was reading a book which I think was the creativity and the age of innovation, something like that he was talking about. The authors talk about, OK, I came across this topic. I didn't understand it that well. I figured, let me go to my department chair and ask if I could teach a

Harpreet: [01:10:44] Class about it. Um, and so

Harpreet: [01:10:47] I try to teach something. You have to learn it. You get to learn it twice. Right. That's kind of like the approach of take with deep learning is I'm learning in public and I'm trying to take a teaching approach to it so that I can try to distill this complex idea, these complex things, and just try to make it as simple as possible for myself to understand and for anybody who comes across my post to to make it easy. Understand. So teach is for entry.

Harpreet: [01:11:11] Yes.

Harpreet: [01:11:11] Enter notes when you need notes.

Harpreet: [01:11:13] Um, that's something

Harpreet: [01:11:15] That I've been stepping up in my gaming review review as much as possible. So review based repetition is a great framework for reviewing. Um, I've got my that cast and set up so that every day just give me a random note

Harpreet: [01:11:29] That I've written

Harpreet: [01:11:30] So I can review that note. Um, yeah, there's some actionable tips there. Um, and then there's another part to the question. Um, but before you do that, marketing insights on how to learn better Mark Rostami tips.

Harpreet: [01:11:42] I mean, I would

Mark: [01:11:43] Echo what's been said already. I think learning in public is huge for a couple of reasons. One is because it allows you to fail and to see that I like to struggle and see that other people share that sentiment and that genuineness is appreciated. And also, there's a more functional thing around that of just like [01:12:00]

Harpreet: [01:12:00] The more you're out there sharing

Mark: [01:12:02] Your journey, there's the more visible you are and the more you can create these network effects just out of out of nothing. And if and if you keep all this internal and to yourself because of this worry of being looked at a certain way, I just know I mean, I've gone through this recently. I've been doing a round of hiring. And it's it's a unique role in the sense that we're asking for folks to create content in addition to technical work. But I really do think from a visibility sense, like sharing your journey is becoming an increasingly valued approach and skill set to different kinds of employers. From what I've seen, just the fact that you have that out there and it gives you an artifact that you can see how you've progressed and it allows you to tell your story better. One of the things I'm realizing is if you if you're just you have a technical skills but you don't have a story attached to it, it's much harder for people to connect with the journey you've made. So I think like having that artifacts and that journey documented is really, really important because it allows you to see how much you've learned over time and it allows you to craft your story, which I think can be really important for for hiring managers, especially the smaller companies, to want someone who can do a bunch of different things. It depends on, obviously, where you're looking to break in. But I think there is there's something to that. And think smart people, good hiring managers, people who want multifaceted people see that as genuine, see it as sincere and want that

Harpreet: [01:13:20] On their teams. Excellent, excellent tips. Um, Marc, if I was

Mark: [01:13:25] I mean, I feel like there's nothing much to add. I mean, I think both comments are really great. I just want to highlight my LinkedIn strategy. It revolves around two pulverize is vulnerability and teaching and learning as I go and like sharing that out. And that has opened up way more opportunities. Kind of like Austin said, like employers find me now because I'm constantly posting what I'm learning and eventually, like, I learn it and I share the lessons learned and the employees feel confident that like, oh, he can bring that to the table for four X, Y, Z role. And so [01:14:00] what you will find is if you start

Harpreet: [01:14:01] Sharing pretty

Mark: [01:14:02] Publicly and pretty. Often you'll you'll build kind of like the network effect, you'll start having people coming to your door and mail saying like, hey, you can't work at X, Y, Z company. And that all comes from just like my content unknown normally start saying, hey, I found you via your post or your comments to seem really interesting. I saw your profile. I would love to chat to see if you want to work here. It happens so often when you put

Harpreet: [01:14:25] Yourself out there.

Harpreet: [01:14:26] Yeah, 100 percent agree. I think those are the funnest types of things

Harpreet: [01:14:29] To do like those

Harpreet: [01:14:31] Candidacies. Sixty six days of Data. There's like twenty one day challenges. And I mean, you put it out there and what's the worst that can happen. Maybe you interpret something incorrectly and

Harpreet: [01:14:41] Somebody might

Harpreet: [01:14:42] Ask a question or somebody might offer a tip. And you learn even more of one thing to not be afraid of is those that, uh, forgot who posted it. I think

Harpreet: [01:14:51] I looked in his name,

Harpreet: [01:14:52] But I'm pretty sure it's the Google chukka. Uh, he talked about the statistics, police on LinkedIn and how when people are sharing, uh, knowledge on LinkedIn, uh, there's like a couple of individuals, particularly in the domain of statistics, they'll come and just, like, tear you apart for whatever reason, ignore those people. Not everyone is like that. Most people want to help you and most people want you to

Harpreet: [01:15:16] Learn and

Harpreet: [01:15:17] Grow it, improve access. Some great comments here in the chat. I go for it,

Mark: [01:15:21] As I say. Quick comment on haters like also is your platform. So you're going to have haters being like, what a you know, like, why are you speaking on this? Like, you don't want those people in your circle and they show up, just block them or move them. You're not as easy to anchor on those people. They they're just bitter that they're not sharing and they don't feel comfortable. Share your stuff, because the more you can gain way more from all the people learning together or maybe be more willing to lend a hand like I learned this year, my tips, then the few people who are going to be naysayers

Harpreet: [01:15:53] At Porter says,

Harpreet: [01:15:55] Remember the three

Harpreet: [01:15:56] Ts?

Harpreet: [01:15:57] Things take time, things take available time. But [01:16:00] what are

Harpreet: [01:16:01] Some of these epic gem comments you're dropping in the

Cristoff: [01:16:05] Scam along with me, along with what Austin was saying about sharing your journey and it's like your journey. I was just saying about how your journey is really like one of the only unique things that you have to offer, because, I mean, that I

Harpreet: [01:16:17] Have to offer because there are a

Cristoff: [01:16:19] Zillion

Harpreet: [01:16:20] Other analysts,

Cristoff: [01:16:22] More or less like me. And and so what do I have to offer? What I can offer? First off, I can get myself a little bit of visibility just by showing and sharing what I'm learning, because on the surface, we all look pretty much the same. It's like it's like you say when you see members of another

Harpreet: [01:16:37] Species, they all look

Cristoff: [01:16:39] The same. It's like even looking at other

Harpreet: [01:16:41] Humans on the surface.

Cristoff: [01:16:43] We all look the same until you can get some insight into who we really are.

Harpreet: [01:16:47] And this is a

Cristoff: [01:16:48] Great way to do it. And that's one of the reasons I like posting like on LinkedIn is because it's like like little snippets,

Harpreet: [01:16:54] Little windows that they last

Cristoff: [01:16:56] Longer. They're not. They last they'll be there next week. They're there in the middle of the

Harpreet: [01:17:00] Night, whatever.

Cristoff: [01:17:02] When I'm when I'm not right there to show and share my journey, somebody can see that. And I used to like people comment on my posts and say like, oh, it's been cool to see

Harpreet: [01:17:11] Your Data journey, your

Cristoff: [01:17:12] Python journey or your school journey and things like that. And you used to kind of bug me

Harpreet: [01:17:16] Because it was like you're

Cristoff: [01:17:18] Saying like, good to see your school journey. Like, I don't know anything.

Harpreet: [01:17:21] Like I know some

Cristoff: [01:17:21] Stuff, like hurts my ego and I want to show that I know something, you know, but then I like actually it it really doesn't matter. Like it is my skill journey here it is wherever, wherever here is it's here, it's now I can't be anywhere else. And if I wait until later, until I'm there to share it like I'll never be there. And so the journey is really all I can share and seems to be good enough.

Harpreet: [01:17:43] Yeah that's absolutely love that man. Yeah. People are

Harpreet: [01:17:46] Not

Harpreet: [01:17:47] Replaceable man. Your your, your you. There's only one. You share your journey. It's not like I can have somebody else pop into @TheArtistsOfDataScience happy hour who happens to be named

Harpreet: [01:17:57] Eric with a tie

Harpreet: [01:17:58] Fighter background. Andrew [01:18:00] gives me gives all the same Beatles and all the same vibe for the happy hour. That won't happen. Only Eric Simms can be happy what it is. Right. I love that. Go for it.

Mark: [01:18:12] Yeah. Mark, you've talked about this quite a bit in different sessions about the importance of being able to communicate with different stakeholders as well in actually in a job, especially the Data job, where the functional functions can be very different and undefined at times and this and that. I think there's this other benefit to sharing. It's like you hone your ability to communicate what you're working on over time. And if you keep it all internal, it's it's just harder to learn how to structure that kind of content, harder to learn, how to structure that kind of communication partner, to learn how to talk about the caveats in your work and where there's uncertainty and where there is more certainty. So I think just like in a functional sense, just like practicing that and then getting the other benefits of building a network, sharing your journey, all this kind of stuff like that's just becoming more and more and more and more important is the ability to communicate this stuff. Because like you were saying, there's just like I was looking at. You get resumes in for a job. Thousands of people can, and there's enough tools and abstractions and piloting frameworks now or just like anybody could go, like anybody can train a regression model, like there's thousands and thousands of people I could find that do that.

Mark: [01:19:18] But the people who separate themselves are the ones who have an identity around it or have something I can see I can ingest. There's a journey. There's like there's an evolution in communication skill that I can actually see when I go look at their LinkedIn. It's not just like a blank page is actually like this whole lineage of this, like Data around what this person has done. So and you can see how it's evolved and you can see that this person is learning and growing and evolving. And that's the kind of person as a hiring, as someone who's been hiring is that's that's the kind of person I want versus someone who can just cotan the model or train a model that does X percentage accuracy on a Data set, like are less interested in that and more interested in like how do you differentiate yourself around those margins? Because it is that last [01:20:00] 10, 15, 20 percent. We've been talking about I think this is part of that. There are other things as well, but this being a significant part of that lingering

Harpreet: [01:20:08] Last five percentage points, I absolutely

Harpreet: [01:20:10] Love that. And that that directly hits on the part of the question that Joel was asking about how do I become more employable like this and everything that that Austin just said?

Harpreet: [01:20:17] That is absolutely on point. There's another question

Harpreet: [01:20:21] You have in there about learning Python. You completely, uh,

Harpreet: [01:20:24] Currently learning you

Harpreet: [01:20:26] Have no professional background in coding, but currently learning Python. Any advice would be great. So if you're not sure we're on the spectrum, you are of the python learning. But my favorite, absolute favorite resource that I've stumbled upon recently that I've been recommending to everyone is python principles. It's it's free. Should still be free. Um, but if it's not for you to like a nominal cost. But here's what I like about it. It's really structured the lessons. This really gets to a super solid foundation and and lets you build the groundwork for learning more. Everything's entirely Web based. So you write your code here. So you have the problem. Same here. You write code here and have your output there and then it'll give you like little tips and stuff.

Harpreet: [01:21:08] And then on

Harpreet: [01:21:09] Top of that, they have these challenges, which are really fun. Um, and these are great to help build your problem solving ability, because these are essentially like word problems like that. You know, how to get math word problems in school, right. Like train leaves be at 10 miles per hour and train X leaves city Y at this. How long until they meet in the middle or whatever? Those type of word problems in math like these types of

Harpreet: [01:21:34] Problems are the programing

Harpreet: [01:21:36] Equivalent of that. You get a lot of reps and do these. These are great because they'll just help you build your confidence, um, before you get into the to the interview mode. But then after that, my favorite book for for a living python is particularly or and if there's a book by W McKinnie Python for data analysis, he's the guy that, um, wrote the [01:22:00] actual panda's package. So no one else, no better person to write that book than him. Um, any other tips on on learning Python.

Harpreet: [01:22:08] Either Mark or I use

Mark: [01:22:10] Dataquest, I teach myself Python similar format. It is a bi worth every single penny. It got me up to speed our new hour, so like our new coding language, but it got me up to speed in Python about a month. And speaking of like, how do you get work experience? My first job out of grad school wasn't and Data role as an operations role. And they use a lot of spreadsheets, Excel or Google Sheets, to do a lot of their work manually. And so Python was a great avenue for me to automate all those processes. And so I got work experience. I forced myself to use Python to do those same tasks, create scripts that took a 10 hour project until like ten minutes. And I just started sharing it with all my colleagues because they want to save time, too. And so now when people look at my operations role, they think of some technical role when in reality only five percent of it was actually using Python. So I find the small, low hanging opportunities at your current role, if possible, where you can actually implement something even just for yourself, but you can practice it.

Harpreet: [01:23:11] Excellent, excellent tips. Um, Eric, any tips there

Harpreet: [01:23:15] For learning Python? But everything

Cristoff: [01:23:17] Everything good?

Harpreet: [01:23:18] Yes, everything's been covered. Oh, well,

Mark: [01:23:22] One more thing. If you have an editor using it. So if you're not using your notebook, if you have an editor download like a lynching tool, like fake eight or or something along those lines, because it'll force you to learn how to write better code. And I wish I did that earlier because I've had to unlearn all these bad habits.

Cristoff: [01:23:41] Yeah. Early on. That's something I wish I'm not I wish I had done differently.

Harpreet: [01:23:44] Set it like it wasn't around

Cristoff: [01:23:45] Then, but like the code with like the Jupiter addin or not Jupiter addin or whatever, like yes, code is so awesome. I'm like so convert it to especially now that you can use Jupiter. Notebook's in it too. It's just like I just, I just work it all [01:24:00] in the same location. So that's like my new face. Ed, for sure,

Harpreet: [01:24:04] The vote count is so nasty,

Harpreet: [01:24:06] So clean as using

Harpreet: [01:24:07] A pie chart for a while, but the Dsdj.co/artists is definitely my absolute favorite. Once I get more, I want to go straight to them. That's when you know you know your oggi.

Harpreet: [01:24:18] But I definitely tax code.

Harpreet: [01:24:20] Um, there's another question there about,

Harpreet: [01:24:22] Um,

Harpreet: [01:24:24] Has anyone tried Tagore's python? Of course, I have not tried Cagle's Python course.

Harpreet: [01:24:29] But, uh, I know a lot of people have been enjoying it.

Harpreet: [01:24:32] I've seen a lot of people posting about it that they've been getting a lot of value from it and learning a lot. So that's probably a good avenue as well. Um, so definitely check that out.

Harpreet: [01:24:39] If, um, if the

Harpreet: [01:24:41] Principal isn't your thing, I have to recommend checking that out. Um, awesome. And, you know, I'm just into another self plug

Harpreet: [01:24:48] If, uh, if

Harpreet: [01:24:49] You guys don't mind, because he's talking about being employable. I happen to be launching a course called the Employable Data Scientist. So check it out, guys. Um, so this is definitely a course that is not for beginning Data scientists. So if you're new to data science, don't take my course because I'm not going to teach you the basics of the fundamentals. This is all about how to think like a data scientist. So how to do projects, how to think through projects, um, how to work like a data scientist, how to set yourself up with project management skills, working and sprints, thinking like a scientist, uh, talking about the scientific method in particular as it relates to data science. I come up with an analysis plan, uh, how to think like an engineer. I we talk about notebook scripts,

Harpreet: [01:25:30] Introduction

Harpreet: [01:25:31] To GitHub and Docker and stuff, and then, um, I think like a business person. And then, yeah, it's, uh, I try to do some completely different from what's already been done. So hopefully you enjoy this.

Harpreet: [01:25:43] Uh, in terms of is that language

Harpreet: [01:25:45] Agnostic for my course. Um, I guess so. Just teaching you how to think like a data scientist.

Harpreet: [01:25:51] Um, so I would say

Harpreet: [01:25:53] This linguistic aspects of look, look, look at that in early September should be Latin. It got a lot of recording to do this [01:26:00] week or next week. Um, does not look like there are any other questions. I guess we can go ahead and wrap it up. Um, the question questions coming in from anywhere. Awesome. I guys will take care, have a good rest of the weekend. Uh, continue to follow. Twenty one days of deep learning. Uh, if you guys got questions or comments, definitely, uh, post them right there. Um, you know, shout out to Nisha. She's had a couple of really good questions. And the thing I like about those questions is that it you know, if I don't know the answer to it makes me go and research it and learn it

Harpreet: [01:26:32] And uncovers

Harpreet: [01:26:33] Things

Harpreet: [01:26:34] That I'm not understanding

Harpreet: [01:26:36] Well or things that didn't explain well. Um, so if you guys are following along with twenty one days of deep learning, definitely, um, you know, let me know what your questions are hosting right there and the comments. I'd really, really appreciate that. Um, I think after 21 days of deep learning, I probably do, uh, 21 days of papers that I'll be fine dissecting a bunch of a bunch of interesting research papers and just philosophy of Data science

Harpreet: [01:26:59] Class of probability to have a

Harpreet: [01:27:01] Paper. It'll be fun. Um, guys, thanks so much for hanging out. Be sure to tune into the podcast release an episode of Jacquelyn Wells, uh, last week was Jonathan Tesser and a bunch of other awesome stuff. Uh, and he says that he's listen to all

Harpreet: [01:27:14] 165 episodes of

Harpreet: [01:27:16] The podcast. Yeah, that's right. Thank you. Thank you so much for listening to to all the episodes. That's awesome, man. I appreciate that. I really appreciate that. Uh, now everybody else go and listen to the remaining one hundred sixty seven episodes. It'll only take you, uh, probably about four hundred hours of of your life. What is it. Four hundred hours well spent. Um, if you listen for forty hours a week, uh, you get ten weeks.

Harpreet: [01:27:41] I guys

Harpreet: [01:27:42] Take cash. Shout out to Matt Bratten. Thank you so much for giving me the shirt man. I really, really, really appreciate it, man. You know, he didn't have to do that, but thank you so much. I appreciate that. Uh, Matt Bratten, uh, and we'll go ahead and wrap it up, guys. And as usual, my friends, remember, you've got one life on this planet. Why not try this on big year, everyone? [01:28:00]