What is a Tech Lead REALLY (Zach Wilson) - KNN Ep.97
Updated: May 20
Zach Wilson is a tech lead at Airbnb and a content creator on LinkedIn with over 150k followers. He is passionate about data engineering and mental health. He also has an awesome YouTube channel, Data with Zach, where he teaches people about data and data engineering. In this episode, Zach explains what it means to be a tech lead, a position that I was personally confused about, how he was able to reach leadership in such a short period of time, and why he finds jumping between engineering and data roles has led to his own personal growth. Zach brings an awesome energy to our conversations, and I’m happy we get to share this talk with you all!
[00:00:00] Zach: And that's what I think one of my strategies that I have on LinkedIn as well is definitely like sometimes I'll write a post and then like, I'll read through it. Okay, this is not what I like. This is just, you know, fluff. If you cut out all the fluff and just get people like the high value ..., like, if you can say what you want to say in three or four sentences, but like you spend three paragraphs saying it like your post is going to bomb, but like, even if it's the same information conveyed, it's just that like one, like you can eat in three sentences. It takes like a, you know, almost a minute to read like that. That makes a huge difference.
[00:00:41] Ken: Zach Wilson is a tech lead at Airbnb and a content creator on LinkedIn with over 150,000 followers. He's passionate about data engineering and mental health. He also has an awesome YouTube channel called Data with Zack, where he teaches people about data and data engineering. In this episode, Zach explains what it means to be a tech lead, a position that I personally was confused about. He also talked to us on how he was able to reach a leadership position in such a short period of time, and why he finds jumping between engineering and data roles has led to his own personal growth. Zach brings awesome energy to our conversations, and I'm happy we get to share all of this with you all.
Zach, thank you so much for coming on the Ken's Nearest Neighbors podcast. I've obviously been acquainted with you on LinkedIn and you're posting some incredible content there. You've started to move a little bit more into YouTube as well, and I'm really excited to see that and to share a lot of those things, you have an incredible story.
You know, moving into a tech lead position at Airbnb, and I'm just so happy that you could come in and tell it, and I'm sure we'll get pretty deep in, into the nitty-gritty of data engineering and quite a few other pretty relevant topic these days. So how are you doing and thank you again for...
[00:01:50] Zach: Thank you. Thank you so much for having me on this podcast. I love making content with people then. I, yeah, like I'm doing really great and I really excited to talk about data, data engineering, all that stuff. That's my passion for sure.
[00:02:05] Ken: Excellent. Well, so I told you this and I tell all of my other guests offline, but for those listening, the only requirement to come on the podcast or that I have as a podcaster, is that the guest is excited about what we're talking about.
And I think Zach definitely checks that box. So it's very, very exciting for me and we're going to have a blast, so let's dive straight into your story. So something that I'm always interested in is how did you first know that you were interested in data? Was there like a pivotal moment? Was there something that happened or has it been a slow progression over time and that you've gotten acquainted with this domain?
[00:02:47] Zach: Yeah. Okay. So I think one of my very first things, first experiences with data was in high school. Like my senior year we did we had a capstone project, right, of like something we had to do that was like involved. It was like, it's supposed to take out the whole quarter and you're supposed to like have this really involved process with it. Right.
And so, you know, a lot of people that, you know, awesome things like, you know, go and work at a charity for awhile and they, you know, change the world or whatever me, my nerdy person, what I did was like, I was, I wanted to do an analysis of: Does my mood impact my Halo play, like, so I was very good at Halo, like Halo, you know, shooter game.
And I wanted to see like how my mood impacted my performance in the game. Right. And I checked it based on a couple of different scales, right. One was like calm to anxious. There was one that was sad to happy. There was one that was like a relaxed to anger. And those were like the three main. And then the last one was energized to tired to energize.
So I had those four that I had to rate my mood. Kind of subjectively before I played each game. And then I would record my stats at the end of the game. So then, you know, have my input and, you know, kind of generate my training set for this to see like, Okay, what's the ideal mood that Zach should be in to be at peak performance in this game.
And I ended up playing like some weird number of games. I think it was like a thousand games. And then like every time my mom was like, Zach, you should quit playing video games. And I'm like, I'm working on my capstone project. Like I need, I need to get more data. I'm just collecting data. That's all I'm doing.
I, you know, I'm not playing video games and just collecting data. Anyways, I did learn that like like the most important one was the tired to energize the bar. Like, you know, if you're tired or sleepy, then like that's, that's the one that had like, was way more predictive than any of them.
But I'm being super angry was also one that was it had a negative impact on the play. Cause like I ended up being more aggressive when I'm super angry and then like, I make bad calls and I rushed you hard. And then two people shoot at me at the same time and I get wrecked. Right. So, yeah, I'd say that, that was kind of like my first kind of, you know, thing with data that I did, that I was like, Wow, this is pretty cool. This is pretty awesome. Like, yeah.
[00:05:12] Ken: That's awesome. Well, that's like the thousand IQ played to say, to figure out a project that you could play video games with and just have the ultimate excuse for your parents. I wish I was that clever.
Incredible stuff. So where did, where did that, you know, where did that take you? You know, after that, you know, you worked on that, did you decide to do something in college related to that? Or was it just a...
[00:05:41] Zach: I was, I mean, that was just like my first experience that I really noticed data as like, Wow, this is a cool thing, but then I kind of like, actually I dabbled a lot in high school and college.
And like, so my first declared major in college was actually pre-med. So I switched from pre-med. Then to chemistry, like organic chemistry and chemistry. It was a big thing. I was like super interested in. And then after that I realized no math is my thing. And then my brother showed me programming.
And then I was like, actually it's math and computer science. So that was kind of like my journey of like switching majors and checking things out and doing a bunch of stuff. I did, it took a lot of classes. That was a big thing I did. Like I was taking like the like 20 credit hours, like every semester.
So, I really kind of soak up information. Right. And that's why ideally landed on like, you know, math and computer science, because they really fit really well together. Initially like in college and everything, like at the end of college, I was actually working on a different project that was kind of unrelated to data, but it was actually is this application for magic, the gathering and what it did was you could build a deck, a magic deck, and then you could play it and you could put like, pretend to play it and see how it draw.
You can draw the cards, you could see like how expensive the deck was, like, both, like, in terms of. In game manner, and also in terms of like us dollars and then you, then you can buy the deck online, get the physical cards shipped to you. That was like my my capstone project for my degree was that application.
And I kept working on it even after that. And I like, so that was such an energizing experience for me that like, I actually believed that I was going to be a mobile. That's why I had a dream of being an Android and iOS mobile developer when I initially graduated college and I was doing a bunch of contracting work and stuff like that.
And then what happened was I met up with these people at this company called red brain labs. And they were like, do data science is the future data sciences where we were, where things are going to go like, and like, and then they were showing me, I'm like, Oh yeah, if you come and work at this company, it's a startup.
It was like 20 people in the startup. 12 of the 20 people have PhDs in math. And I was like, Oh, that sounds like a pretty smart company. It sounds like that might be where I want to work. And that's where I like really kind of dove really had burst of the data. Was that like, I'm still doing like Android and iOS dev on the side, like for contracting and stuff like that.
Like just to make a little bit of extra money, because I was like an intern you know, doing data science, cause like the mobile, the mobile stuff paid way better than being a data science intern. Right. It was like five, five times better. Right? It was, it was a very big difference.
And so I did that, like make a little bit extra money and then I still just kind of focus more on data science and yeah. And that's like, where I kind of like really launched my career was there. My first job I like did a lot with SQL just like tons of SQL. SQL and like Vertica, or I've heard them Vertica Vertica is like a database, it's like kind of like Teradata and like the other kind of analytical, like Oracle kind of those analytical kind of you know, old school-ish databases and yeah.
And like, I did that. And then I worked with Tableau, lots of Tableau. I so much Tableau that I became a Tableau Certified Professional I'm legit certified. Right. But yeah, like I made some very fancy dashboards. It was fun, dude. That was kinda like, that's kinda where I kind of broke into data and like really like made a career of it at that point.
[00:09:17] Ken: I do want to rewind a little bit to your college.
I love the project. I mean, it seems like the two projects that you are taking related, there were very intimately tied to your interests. I think that that's something that I think is really compelling and it seems like you had an infinite energy to collect samples and to be able to work on those. I mean, so I dabbled with some magic, the gathering back in the day as well.
I even like maybe it was last year. A couple of years ago, I tried the Magic the gathering arena. They have, they have like a new web platform now, which is which isn't that bad. And that to me is like, it'd be very interesting to pursue some sort of model to, to actually play that game. Right. We focused on closed games.
We have go, we have a lot of these things, but these open-ended like, choose your deck, choose these things. Cool complex problems. I mean, a lower hanging fruit, but still arrive on one it's like Pokemon and like figuring out can an AI optimized on a game like that can completely tangential. How long do you think we are away from being able to have someone or like an AI that would be competitive in that?
I mean, we can do, we can do StarCraft. We can do these types of things, but there's so many less moving parts than in a completely open-ended card game or something like that.
[00:10:44] Zach: I think, I think there's a, there's a couple of things that are like cause they're one of the things that I think is interesting about the card games and stuff like that as well is that there is more of like an orangy into it.
Right. So I think in some way, some regards, I don't think that it will be possible for AI to ever be as dominant as they are, as it is in like chess. Right. Or as it is, and like other things, because. S almost completely like 100% skill, right. There's almost no randomness in chess at all. Whereas like car games have that like shuffle back randomness and like is sometimes if your opponent just draws the right eight cards at the beginning of a magic game, like you're done, you're done, dude.
You're done. Right. It really depends on like how, like the decks kind of play out. So, but like, I think that. There, there is some decent AIs now, like in like some of the magic games and you, like, you can play against like on like the online magic games. Right. You can play against AI's that are, that are decent.
But I haven't like seen one that was like super dominant though. Like, yeah, it's a tricky one. Like one of the things I did when I was building out that. So that app actually got pretty successful. Like I got it got up to about 200,000 downloads. So like I had like a lot of like decks, right.
Because like, then what they would do is they'd build a deck in the app and then that data would then also be stored in like my MongoDB database. And then, so, and then I would export all that data into like SQL and like kind of more of like a place where I can do more of like the relationship modeling and I actually built like a recommender.
So what it would do is like It's how I would kind of build like the, like, people also saw this card. So like, if you viewed, like in the car detail, in the app, there would be a little carousel at the bottom that has like cards that are that are paired with this card. Kind of like a wine pairing thing, right.
Where it's like every card, like you look at the decks so that a card is in and then you see what, what, what cards also show up the most in those. For that card, like excluding really, you know, common things like land and stuff like that, but like Mo like the rest of the cards. Right. And that was a really cool way to like, kind of build like a basic recommender.
Right? No statistics, no regression. It was all just a recommendation. So. Based on counts. Right. And like, it did pretty well. Like I honestly thought the recommendations were pretty good. So yeah, that was, that was fun. And that was, I kind of dabbled in like the AI ML stuff there as well. That was, it was a fun little project as well for sure.
[00:13:15] Ken: That's awesome. ... crowds is, is always pretty compelling. And we also, I don't know, like a super interesting space. That's one of those where. I feel like I'd want to go down that rabbit hole, but I would get completely lost and not want to do anything else for awhile and put that off for a later day.
You know, I am interested in sort of how you've made. That first career transition into data. So you had the internship and you, you talked a little bit about, you know, sort of what your, your first job was like, but what was that progress process like kind of breaking in and and, eventually I, did you keep doing the iOS development or the Android development, or did that stop...
[00:14:01] Zach: So that's a great question. So. The internship was like three months. And then I got a full-time offer from the same company. And then I stayed there for about another nine months. The reason why I left that company was they actually got acquired. So they were like a little startup called red brain labs.
And then they got acquired by Savvysherpa and Savvysherpa is actually a subsidiary of UnitedHealth Group. And so like after like the startup culture kind of disappeared, I was like, I don't want to work here anymore. This company is, it's not as inspiring as it was when I started. And so then I went and I got a job at Teradata doing data engineering, right.
Doing like doing like big data and stuff like that. A big thing that was important for me back then was knowing I, and I had to do a lot of studying and research on this stuff. It was, I had. About Linux. Linux is super important. Like Linux and Unix. Was one of the reasons why I like coding on a Mac, right.
Is because Mac and Linux are like, the command line is like the same for the most part. Right. And so you get the benefit. If you're on a Mac, you just getting Linux for free, but if you're on windows, it's like, Oh, it's a completely different set of commands at the terminal right.
[00:15:10] Ken: Now with WSL2, Linux has been very easy to use on the Windows machines. I highly recommend it.
[00:15:18] Zach: Yeah. I switched in 2014. I switched when I graduated college from windows to Mac and I have not looked back. It's been like eight years and I'm like, sure, apple, here's another, like, whatever ridiculous sum of money for the next month. But like so Linux is important. Another thing was like Java.
One of the things that was really nice was the, I was very good at Android development, right? Yeah. Android development is also in Java, right? So doing Android development and then like Java MapReduce on Hadoop and big data. Those, all those big data technologies are all like, it's all based on Java.
Right. And so that was made it easier for me to like, get into that new space because I already knew the language. And I was able to, you know, ace the coding interview. Coding interview was like really easy. Like I it was like one of those like easy fizzbuzz ones and they just wanted to hear my thought process.
It was like it was so easy. And then the Linux questions where the hall was the hardest section for me, like where I felt like I failed and I actually thought I didn't get the job. And I was like, Oh, this sucks. Like, I guess I'm still at Savvysherpa for a little bit, but then they call me and they're like, yeah, we want you to be hired.
You. Yeah Teradata was great. Like I kept doing so like when I went from red brain labs to Teradata, so I read Brain labs. I was making about like $50,000 a year in Utah. Right. And then when I moved to Teredata, it went from 50 to eight, 50,000 to 80,000 hours. And I'm like, at the time I was like, I'm a big baller, dude.
This is amazing. I can't believe I got that much of a race. That was crazy. And then I worked at Teradata for a while. But then I realized I didn't want to live in Utah anymore. I got sick of Utah. Utah was like the problem. And then I was like, I need to move. Right. And then I've worked at Teradata for like seven months, which was like, it was like my shortest time.
My shortest time was on Teradata, like of all my companies. And then I moved to DC to work at research. And this company called research innovations, which was just like, I was just doing back end to there. I wasn't doing data. I was just doing like back end development. A bit of full-stack, but mostly backend development.
That's what I did at that company. Yeah. And yeah, it's kind of crazy. And then I got the job at Facebook after that, that was kind of like the kind of transition to Facebook. I've had this kind of pattern in my career where I've kind of like kind of swung back and forth between like traditional software engineering, like mobile dev backend devil's dot gov, and then kind of swaying back more into like data like data science, data, ML, and then kind of swaying back.
Right. I've done that like so many times, like, cause like Teredata was like data and research innovations with software in Facebook was data and netflix was software end. Right. And now Airbnb is data in, right. I've done. I've done the hop. Like it's been freaking insane, like, but also awesome. Like it's kept me fresh. Kept me nimble, you know?
[00:18:11] Ken: Yeah. I was going to ask, why do you think that is? And I mean like, you know, you say it's a pattern, but is there some design in that pattern? I constantly find myself trying to, with my work and with content and consulting and stuff. I try to find myself like working on different things across the platforms.
Like if I have to do a lot of coding for my YouTube videos and I'm doing a project there, I don't really want to be coding in my like nine to five jobs as much that right. I'm trying to outsource, I'm trying to do more client facing stuff and vice versa when I'm doing we're talking head you know, is it something that like, Hey, you get, it can get a little overwhelming and like you get to work some different muscles and then you come back.
Or is it just like, Hey, these are how the opportunities lined up. Like with software engineering, maybe there's more appetite for software engineers who have really good data background. And when you go back to data engineering, maybe there's more appetite for data engineers who are pretty up-to-date with their software engineering.
[00:19:15] Zach: Oh, yeah. And that's a big thing at Airbnb actually. So like Airbnb actually doesn't have that many data engineering compared to like other companies. But like they really hire for like data engineers who have strong software engineering skills because like one, like all the pipelines are right or installer.
Right. We, we ride scholar code at Airbnb for our data pipelines. Right. Which is like, you should look into a lot of other companies that's very different. Right. And we also have unit tests and integration. And all that stuff, which is another thing that is kind of different from especially from Facebook, but also I'm a bit different from Netflix as well.
Like there's an even higher degree of like software engineering rigor on the data pipelines at Airbnb that I've noticed for sure it gets that's one of the things that's so awesome about my role right now, right. Is that I kind of get to do both. Right. And that's one of the things that's like, I feel like that's why this role fits really well for me. Right. And I feel very fulfilled in that regard here right now. It's pretty cool.
[00:20:16] Ken: Excellent. Well, I think it would be too silly of a question to ask what the difference between a software engineer and a data engineer is, but I think it would be practical to say, what could. Maybe most data engineers learn from software engineering and possibly vice versa. What do you think of that?
[00:20:36] Zach: That's a great question, dude. Really good questions there, man. Okay. So on the data engineering side, right? I think that data engineers, where they can learn from software engineers, right. Is a big thing is around like having good, like on-call processes and how to do better maintenance of stuff.
Right. And what was cause sometimes like that's one of the things that's kind of fundamentally different between data engine, software engineers, software engineers, generally. Online systems that like, when they fail, they need it, like you to actually go and respond pretty quickly. Cause it's like down in the servers down and someone's expecting to use the thing.
Right. But in data pipelines, it's like, Oh, the data is just delayed. Right. It's delayed by like a certain X number of hours or whatever. Right. And so like the data and data engineers generally don't care about like having good, solid on-call processes as much. But I, for me, I think that that's actually one of the most important things that a data engineer can do is hit SLA as much as they possibly can.
Right. And a big part of that is managing on-call and how like, Okay, what do we do when things break and then like other things. For data quality checks, right. It's like, Okay, say it barely fails. Say it's barely outside the threshold. Do you just let it, like you just skip the check and let it run or do you have to escalate it?
Like how do you troubleshoot data quality bugs. Right. And kind of like that whole rigor and process around on-call and kind of managing that stuff is something that like, I don't think that data engineers like generally need to do as much, but like if they do it like that, that's how you really get to like a really high level.
Engineering excellence. Right. I think on the flip side, like some things that software engineers could learn from beta engineers, right. A couple of them are like one of those around like efficient data modeling. Right. And how, like really trying to compact that data as much as you can. And also.
Not just that, but also like usable data models. Right. And understanding how to like write data. That's easy to query and write data that like is easy to move around or do whatever you need to do with it. And like, I think that that kind of modeling and that kind of stuff, modeling and query patterns is something that I think data engineers are generally really good at, but software engineers write pipelines to write and they need to write data out and stuff.
And I've just seen a lot of times in those cases where I've like taken over a pipeline that was written by a software engineer. They write everything like, but then like it's but the data is like inefficient or like they didn't take the right or they didn't leverage spark the right way. Like maybe they should have written more or more files or fewer files, or maybe they needed to tune the spark job a little bit, things like that as well, that are like just, you know, nitty gritty technical things as well. But I'd say that's like, those are the things that I think they can learn from each other.
[00:23:15] Ken: So I have sort of a follow-up question. You know, at different organizations, they organize teams in different ways. So I've worked in and consulted in places where data teams were completely separate from software engineering teams.
I've worked at places where they were both under the same umbrella and, you know, like the software engineers, data engineers, data scientists, we're all effectively working in the same group and on the same projects under the same person...
[00:23:46] Zach: I've been in both teams to like like Facebook and Netflix have it.
They're more like centralized, right? Where like at Facebook I was on a team of just data engineers. And then at a Netflix, I was on a team of just data people like Des and Des, essentially we're all on the same team, but and then that was it. And then Airbnb. Like we're just, we're, we're part of edge, right?
We're mostly part of an dry. And so that's where it can be a little bit tricky for like, in the, kind of the more embedded model that you're talking about can be kind of a little bit more tricky to do some of this stuff, for sure.
[00:24:17] Ken: What were some of I don't want, you don't have to go too in depth or something like the benefits and drawbacks of both.
I think obviously they're two different models and there's going to be complexity. Also relationships are very different if you're like throwing something over a fence to, to the engineering teams versus having to sit next to the person who is working on the product. Right after you do or like in parallel with them, you know, I, that, that to me is so fascinating because most people have not had that experience of working in teams that are structured differently like you and I have.
[00:24:57] Zach: Yeah. I mean, this is great. I actually talked about like, in my Ukraine webinar, I actually spent like five, 10 minutes talking about this exact, what we're talking about right here. So I think like when I think back on like the pros of being on a centralized team, like there's a couple of them, one is that robustness right?
Where you can really like for example, like you're only on call maybe like once a quarter, like if you have a big team, right. Or like, depending on how centralized it is that's a big thing that can help is like you have this kind of diffusion of like that maintenance, the maintenance responsibility gets kind of distributed.
Which is great. And like, you can have a lot of really good things there, but you're totally right. Like throwing things over the fence. Cause like then you have to work with a manager and like prioritization, right? And then like a specific product teams might have to wait. They might have to wait three, six months depending on how we help you, how the centralized team ends up prioritizing things.
Right. But in the embedded model, they can plan more effectively because they are just with the, in, in the embedded team sitting with them, like you said, and so that's kind of like a trade off. There is like the prioritization complexities a little bit different. I think like the centralized versus.
I think on the flip side, though, for embedded kind of the minuses, there is like a, you have to train people. You really do. Like, I mean, like for me, I'm on a team that doesn't have very many data engineers, but we have a lot of backend engineers and like some other, like, you know like ML engineers and stuff like that.
And like, we want to. We should all be able to, you know unblock a pipeline. Right. And be part of that maintenance and diffuse the maintenance across. Right. Just because it's a pipeline doesn't mean it has to be restarted by a data engineer. Right. I think that like, but that costs like most data engineers know how to do that just because of like, experience and they've been in that boat and like sometimes those other roles haven't and so you have to like teach them how to do on-call, but like it's worth it.
It's definitely worth it. Like And I think that another thing with the embedded model that can be trickier is like, it feels like I'm getting the right, like knowledge sharing and career support and stuff like that. Like, because working with people more senior than you, like, that was what happened.
And that's what, that's what I did at Facebook. And like, it was amazing. It was so amazing being able to work with people who were just like better than me, but if you're the only data engineer on the team, How do you work with people who are better than you at data engineering, right. You can't right. And so I'm like, I think that that's kind of a, one of the things that you really need to work with your manager on to like, understand how you can grow your career better. So, yeah, so those are kind of like the pros and cons embedded versus centralized, you know.
[00:27:36] Ken: I love that. So something that I see where embedded, I wouldn't say it breaks down, but it becomes a little bit more difficult. So I think software engineering and data engineering. And even like the other end of the spectrum, like ML engineering, those can generally be run very effectively or managed in a similar way.
Once you got, you know, once you start working with data analysts, data scientists, where the nature of the work can be a little bit more exploratory. I think project management system. Can still be effective, but they break down a little bit because the nature of the work is a little bit different. Within an embedded team.
How do you manage those slightly different expectations are slightly different styles of work, because I mean, then, you know, there's might not be a correct answer to that because that's a problem. I see. Almost everywhere. But you know, do you have any thoughts on that? Or, you know, some parts should be holy in project management or some parts should just be run different?
[00:28:39] Zach: I think, well, I think that there's one thing that I should get support. There's a lot of teams out there who think that like, data scientists really need to be a part of like daily stand-ups. So you know how like stand-ups and agile and software engineering, like those things all go hand in hand, like that's like a very common pattern in that, like is usually pretty effective in software engineering, because it's like such more of like a builder pattern.
Right. And like you're creating and building infrastructure and stuff like that. So it's a little bit more predictive. But you're totally right. And like the data science and like exploratory spaces, like the the iteration cycles are just longer as well. Right. So for my team, what we do is like, we kind of do a middle ground where we just have like a weekly check-in or we just check in once a week to talk about the progress of stuff.
Right. So that, like, it's not like as demanding. Right. That's pretty, that, that helps a lot, like trying to meet somewhere in the middle. Right. And like, assuming that like That there's not as many blockers for the software engineers. And in those cases, I work with them on more like. Be better. Async right.
And top talk these through over slack. Cause a lot of times where you can, Slack's powerful man. And you can really do a lot of async goodness there. Right? Minimum minimizing meetings. I'm all about that. All about that, for sure. But like I think that it's definitely tricky. I think if your company doesn't have good career ladders defined as well, if everyone is like on the same team and they all have the same kind of expectations that it's going to be bad to.
Right. It's tricky for sure. I've never been a manager like that where like I've had to manage like multiple different types of people that like yeah, for sure. That's I don't feel like I'm qualified to, you know, have an answer to like...
[00:30:18] Ken: Totally a farewell, you know, that brings up sort of your current work and your role and what the expectations are. And so you're in the tech lead position. Can you walk me through what it means to be a tech lead? Like what are some of the responsibilities and like, how does someone like aim for a role like that.
[00:30:35] Zach: Yeah, for sure. For sure. So for me one of the big things I try to work with my manager on is like having like a 50/50 split between like productive, like creative work and 50% working on like unblocking people and, you know, prioritization and planning and design and all that stuff.
And so I'm like, that's a big thing. I actually I feel like I initially got into the tech lead role in my second year at Netflix when I was working there, they don't have like titles. They're really, everyone's just a senior software engineer, but the responsibilities that I had then are similar to the ones I have now.
So like, it feels like similar in that regard, but I think a couple of things about tech lead is so I will work on like prioritization, right? So one of the big differences between tech lead and like senior engineer is. You you're you're picking which projects should be done first. Right. And the order and like how we actually spend the engineering resources.
Right. And taking the risks and moving things around and negotiating with people and trying to figure out like, Okay, how do we get this request in? And what do we give up if we take, if we take on this, because we only have so many hours to deliver on stuff There's a lot of that. Right. and that takes just a lot of effort.
It's like more of the managerial side of it. Right. I also feel that lead time, like a responsibility to mentor people and grow people and define best practices. Right. And have a way of like, being like, all right here is the best way for us to write a data pipeline or here's the best way that we should do this or that.
Right. And like kind of have good best practices and get those adopted throughout you know, your whole org or like whoever you're working with. That's another thing. I think that's kind of a difference. And then on the technical side I think that the big thing is like you take on the most tech leads usually take on the most complex and risky technical projects.
Right. And so that they can be more likely to deliver know strong value there. That's yeah, that's kind of, kind of how it works. Like some weeks I, you know, I shoot for 50/50 manager and a technical, but like it's never like that some weeks are like almost a hundred percent manager. Some weeks are a hundred percent technical.
Right. And that's like, when it's like that, I'm always like, Oh, that's too much on either side. It's too much. Right. I know I do like it when it's more like 50/50, but you know, you gotta roll with the punches sometimes, right?
[00:32:58] Ken: Oh, yeah. Well, you know, the funny thing is I always try to look at things in terms. Longer chunks of time. And so you're like, Oh, you know, if half the days I spend doing the managerial and half the days I spend doing the tactical, technically at all at nuts, out to 50/50, if I stretch the timeframe bigger and bigger.
[00:33:16] Zach: Yeah. As long. Yeah. Like once you have enough samples of days, right. That's the, that's the expected value, right? Yeah.
[00:33:24] Ken: Exactly. So, that I think a lot of people are maybe confused about just by the nature of the title. Is, is there one tech lead or they're tech leads for different verticals or are there tech leads for different you know, products or, you know, how does that role break down? I mean, obviously that's different for different companies, but what does that mean?
[00:33:49] Zach: Yeah, for sure. So I can, I can kind of like explain like kind of how my role works and everything. So I work as a tech lead for a team called Commercial products, which is actually it's an org of like over a hundred people.
And and so for me, why I'm like, why I'm a tech lead is I manage the data quality for the hoard. That's like, and it's like so one of the things that's important there is it's like a big. Big scope, right? Of like usually multiple teams, right. Where that are involved to like manage these kind of things.
At least in big tech, like tech lead, technical lead in like other cases can be like just working with one team to build maybe a complex project. Right. Because there's essentially two types of tech leads. Right. You have like adapt tech lead because you can also have a tech lead. I'm really good at machine learning or computer vision or some very specialist kind of area.
And they, and they have a very strong depth, right? You can also have tech leads that have very strong breads that are very good at unblocking people and connecting a bunch of teams and integrating APIs and getting all that stuff working. So it's one of the things that's different about how you grow after a scene after being a senior engineer.
It's a lot more like choose your own adventure. Right. It's a lot more, very like what, what, what do you think you're good at? Like, do you want to be more like wide or depth and like how do you want to bet on your skills? Right. I think for me, I'm more of like a brat, like I kind of a breath tech lead in terms of just like, cause I have that.
Skillset and the data engineering skill set. And then I work in data quality, which is like, kind of like touches a bunch of things. Right. It's like, Okay, logging is one piece of it, which is going to be like front end and mobile stuff. Right. And then we have like all the different data pipeline pieces, and it's like, it touches a lot of things as opposed to being like, I'm very good at like random forest and making a prediction on data right.
Where it's like, that's like a little bit more contained in like but, but has a lot of depth to it as well. You know? So yeah. It's like, I think that, that, that's kind of like the two types of tech leads and how they kind of work. Yeah.
[00:35:53] Ken: So, I mean, in like very, probably overly simplistic terms, I mean, do you think a suitable description of a tech lead is like maybe a part engineer part like program manager or product owner.
You're kind of, you know, you're, I think you obviously have significantly more hands-on knowledge than like a product manager or like a project manager, but you still have that. You're still doing a lot of the prioritization. You're still. Guiding vision and managing people in a certain way. It's a very, very cool. And it's a little bit different, difficult to, to conceptualize, I think for a lot of people.
[00:36:36] Zach: Oh yeah, for sure. For sure. It's definitely it's an interesting role and I've I've really been liking it though, so for sure.
[00:36:44] Ken: So how would someone who wants to become a tech lead? How would they like angle for a position like that?
[00:36:51] Zach: Okay. I think the first on there is like to really get good at like delivering like complex software projects. Right? Like if you can deliver a strong value and big value on a complex piece of software, and you can do that, maybe consistently over a little. That's how you really can get that kind of foundation.
That's ready to like potentially be a technique. Then like once you have that, that's kind of like the software senior software skillset. Like, especially as you get kind of deeper into your senior career, you should be like also mentoring people and growing people and teaching people and kind of sending the ladder down and like getting better at those soft skills and those mentoring skills and those upskilling of people skills and Because that will make you a way better tech lead as well.
And once you have that kind of experience and you have like the track record for of delivery, then it's pretty straightforward to like actually get the tech lead role. You can actually land it after that. The depending on like things you still gotta ask, like DSA questions. Right. And you're going to be definitely leak coded up.
Right. I leaked out in before my Airbnb interview. I mean, I didn't leave out a ton then, like, it was like maybe like I spent like 10 hours on it. Right. Because I don't, I'm also not an advocate for like, you know, lead code, all the questions or whatever, you know, that's, that's too much. But I think that Those are kind of the big things, right?
Is you need to get those soft skills, right. Managing people and managing expectations prioritization, because sometimes you're gonna need to tell someone that they're broad, their project is going to have to wait, stuff like that and getting good at that. Stuff's important for sure.
[00:38:28] Ken: So, I mean, you, you touched on this a little bit, but how, how do you go about building those soft skills as much as possible? So just reps, experience, talking with people, or is it something beyond that.
[00:38:45] Zach: Yeah. That's a great question. I think a couple of things, there are a big part of it is reps. And like I'm talking to a bunch of people. Like one of the things for me is that like, I really like to try to build relationships with the kind of a bunch of different people, or I try to understand people's emotions and just really interacting with humans.
Right. And kind of do the human thing. Right. And I think that that's a big thing. Like raps is a big one. I think another one. Like taking risks, right? Because like, sometimes like you want to get in those reps of like communication when it does count. Right. And when, like, when it will make.
And so if you can take risks in those situations to understand things right. And to, or until like, it'd be able to communicate those risks in a way that's effective. Right. And kind of get used to a little bit more like higher stakes communication. Right. And understanding like what your words do and like how they, how they, you know, change things and influence things and impact things.
That is like another thing that I think is important and that's, that does come with wraps, but it also comes with. It's reps, but you can't just keep, you can't just keep repping the bar. Right. You have to, like, I got to put like a little bit more weight on the sides of it to like, be able to kind of keep communicating well in, in situations that have a bit more pressure and a bit more pressure, you know.
[00:40:03] Ken: I love that. I think you're forgetting the most important one, which is to listen to podcasts fairly weekly.
[00:40:10] Zach: Yes, for sure.
[00:40:14] Ken: No, I, you know, I love that, that, that to me is so fast. And thank you for sharing those types of things. I mean, something you're also, you're pretty well known for is also sharing a lot of information that you've learned.
And from your experiences on LinkedIn and now on YouTube, I'd love to hear about, you know, transitioning a little bit from your work, but more about like building a brand and telling stories and like making, creating value publicly, you know, what is your philosophy on that?
[00:40:45] Zach: Well, that's, that's a good one. So I think there's a couple of things in there. Building a brand has been verse off. It's been really amazing. I highly recommend it to most people. I've been doing it for about a year-ish now. I really started in like about February. Like essentially right when I started at Airbnb was when I really started to, you know, be very consistent on LinkedIn and like posts.
I think that one of the things is, is about habits, right? Because habits are like super important to build. Cause like for me, I've actually been. Creating content for a very long time. Like I've been posting almost daily Facebook posts since about 2009. So I've been doing this for like almost a decade in terms of just like writing and just like writing my thoughts for the day.
Right. I wouldn't have those 2009 posts. I'm I was like, Zach, you were an idiot. Like. Like I, so I've, I've, I've really liked to write for a very long time. I, that, so that's been, for me, that's been one of the most important pieces I think of this brand journey has been writing things that people like are interested in and that they get valued from.
Right. And so I think that, like, that's a big thing that I think is important of like that stuff. That's why YouTube has actually been challenging for me, right. Is because. It, it hasn't, it hasn't felt like it hasn't felt like it fit like a glove, like LinkedIn. Right. Because like I've been practicing writing for freaking so long.
Right. And video is just like different, it's a different animal. It's just a different beast completely. And I'm learning though. I feel like, I feel like as like keep practicing, I'm going to get better at that stuff. It's interesting for sure. Like.
[00:42:30] Ken: Well, I love the habituation conversation. And even the story of revisiting your old posts. I mean, to me, I think part of content should be for yourself and for retrospection and looking back and saying, Hey, this is a catalog of all the things that I've been learning and. I can go back and see that if I ever need a reminder of who I am or the things that I picked up, but the other added benefit is other people can see, see this and pick something up from my experiences as well.
I mean, I love podcasting as a platform because selfishly I get. Awesome guests like you, whatever questions that I have about, you know, like your experiences and your work, and I get to learn, but everyone else gets to tune in and listen and share the experience as well. And, you know, there's just this really unique, double, like internal and external benefit that you get from producing in this way.
And you know, a lot of people think I want to make sure that people realize that, like it isn't all. For creating value, right? Like that is really important. No one will watch your content if value isn't created, but it's totally fine. If the motivation for making content is partly for you, because your motivation will be higher.
Like just like your projects right below and magic, the gathering, right? Like you're motivated to do those things because you actually care about the outcomes, you know, if your time posting every day, because I care about. You know, reminding myself of my story or sharing what I'm working on and I get happiness or excitement, or I feel accepted because of those things.
That's a perfectly good reason to go out and produce and create these habits. And I think that that gets overlooked a lot. A lot of people are like, Oh, it's such a slog. It's such a grind to produce things. And I'm like, no, if you are producing the content cause you, you like it. Or like, I have fun when I make my videos, for sure. Like, it's a lot easier to do it that way.
[00:44:34] Zach: Oh, yeah. Well, you gotta be playful with stuff, dude. And you kind of channel that inner child and like just do right. Things like for the sake of writing them and like realizing that like, you've been like a lot of times, if that, that, that also is kind of a secret sauce as well.
Right? Because like, if you are playful and you're are having fun with it, sometimes you write better content because it's more sincere. Right. And it's more like, Oh yeah, I'm speaking from a place of like who I truly am, not, Oh, I'm professional Zack. I wear a suit and tie and like, Oh yeah, I just got into Google.
I know like some like automaton robot guy, right. That like, there's a lot of posts on LinkedIn. Like that, that just bomb. And they don't, they don't go anywhere because they're, so it was like, so young. So like get out of here, dude. Like I get that. You want to be like a, you want to present yourself as like professional, but like you have no flavor.
Like you have no, there's nothing going on here, dude. This is all bland. Right. And I think that, like, that kind of like building of your own flavor, like, at least for me, that kind of playfulness of like building your own flavor is like, It's great because like, you can refine it and it's a craft and art and you can get better at it.
You can learn and grow. And like, I definitely know I've gotten like, Committing to right. Because last year I posted 700 times on LinkedIn. Right. And so like, that's a lot of posts, right. That was like, you know, it was almost two a day. Right. And so like, that was, I learned a lot from that, just the, those, those habits and that consistency, you know, and that's, and I feel so good about it though. Cause now that I can write, I know I can like, just get into people's heads and change the world. Right? Like it's awful actually.
[00:46:15] Ken: I think a lot of people overlook the importance of writing, especially post pandemic. I would argue the main way we communicate with people now for the majority of people in technical careers is through, through writing, not as much through talking, right.
You do stand up in the morning and then the rest of the day are communicating. Yes. For the most part or one of the other platforms or, or email or whatever it is, and to be able to convey information very clearly and get people to read it and understand what you're saying, unbelievably valuable skill.
And you know, if I'm hiring someone, I look at blog content that they write, can they, can this person write if that's how I'm gonna be communicating with them most of the time, like it's pretty cool to be able to also showcase that in any of the content you create and. Get better at that as a medium in and of itself.
[00:47:07] Zach: Yup. It's and that's, I think one of my strategies that I have on LinkedIn as well is definitely like sometimes I'll write a post and then like I'll read through it. I am like, Okay, this is not what I like. This is just fluff. And then if you cut out all the fluff and just get people like the high value ..., like, if you can say what you want to say in three or four sentences, but like you spend three paragraphs saying it like your post is going to bomb, but like, even if it's the same information conveyed, it's just that like one, like you can eat in three sentences.
Another one takes like, you know, almost a minute to read like that. That makes a huge difference. Right. And like, conveying things like succinctly and like just getting into people's heads. Like that's you got to respect people's time too. That's one of the things I think a lot of creators forget is that like, they feel that like the algorithm means that they're entitled to engagement and end to end to reach.
Right. Cause a lot, you know, you see these creators on LinkedIn that are like, Oh no, the algorithm changed. And now my content doesn't do as well. And it's like like it could be like quality reasons for that, but also like you have region your, your audience gets to decide that right. Cause highly engaging content and content that people like and love is always going to do well.
That's just how it works. It's how content works. Right. And like an hour, if an hour, like if your content strategy is so shaky that like an algorithm change, just like derailed. Then you need to rethink your strategy. Cause there's probably something you're missing or something that you could do as well to, you know, make your content better. So I it's definitely, it's an interesting journey for sure though.
[00:48:45] Ken: Well, the time sensitivity is really interesting for two reasons related to YouTube is that YouTube doesn't value the user side. Right? So YouTube, I, my goal is to get the person to watch as long as possible. Right. It's not to get conveyed the information and as concise and digestible away as possible.
The exact opposite of what someone learns in business school or in any of these places. It's like, Hey, get to the point. Don't waste. Anyone's time, like start with the important details. And I personally find that very challenging when creating content, because I just want to be like, Hey, like, these are the things you should know.
This is why, but like, you really don't have to keep going on, but it's difficult to continue to like perpetuate your content with the algorithm. If you're not playing the game of telling like long convoluted stories. And so finding, finding the art in that of being able to, to like tell a longer story and make it compelling and make people want to follow along is something that's been very novel and a little bit difficult for me because I like to get to the point and.
Yeah. I find that very interesting. Like, you know, that might be one of the reasons why you're like, man, YouTube is it's harder for me to motivate because you have to tell a story in like a very inefficient way. Right?
[00:50:03] Zach: Yeah, no, definitely. Definitely.
[00:50:08] Ken: But yeah, I mean, so one of the things I am interested in, I mean, obviously there's been just consistency. You've made 700 posts. You've put yourself out there a lot. A lot of people struggle with the risk-taking element that you described before. How do you experiment, how do you whether it's with interpersonal communication with a post or whatever it might be, how do you take those calculated risks without going too far?
[00:50:35] Zach: That's a good one. That's a very good one. I think there's a couple of things there. Like I think one piece of it, I, when, when I'm writing content, I think like, Okay, who am I really going to help here? And like, is this gonna, is this gonna help them enough that like it's worth the risk. Right. And it's worth the and some of it, like I've realized that like, there's a lot of things.
I think are risky, but actually aren't risky. I mean, like, I've definitely made mistakes. I've made big mistakes in my content and I've taken risks that were too far. I've got like, those definitely I've deleted a lot of posts. So I probably deleted like a hundred posts. Cause I'm like, no, this is like for a couple of reasons, like either like it's a low quality post or.
I like I write it and then I'm like, actually, like, I don't think I actually believe that that's just like, I'm just writing right now for like engagement or I'm just writing for whatever reason. Right. And a couple of times last year. Right. I actually landed on like LinkedIn lunatics. Right. And I'm like, yeah I need to be better.
Like I need to like know and I'm like, and rightfully so, honestly, rightfully so. Like I give them, I give like, honestly, like the internet, like it was like, no, Zach you're being dumb. I had to think about it for a little bit. And like, and I'm like, yeah, internet, actually, you win this round. Like, I messed up that that concept was too far.
And so, so like it's tricky. Right. And I've learned and that you get feedback, right? You get feedback pretty quickly. Like if your content is like, I know too controversial or too. On the line or whatever, or like, and you get it two ways you either get like internet hate from it. Right. You get people who post your shit on LinkedIn ...
And then they like, you know, you get it, you get hate or you or you get no engagement. Right. Or it's just bad. Right. It just dies. So it's either. Those are the ways that I recognize that like the risk didn't pay off. Right. It's like, you're going to get stuff in one of those two buckets, but there's a third bucket, which is just like, there's also just like generic haters on the internet as well.
That like, just like composts to like take you down and tear you down. Those people don't listen to those people. Like really just block those people. Right. Cause like there's a lot, there are people out there. There are feedback isn't worth anything, right. Their feedback is not even feedback. It's like a putdown. Right. It's just like you know, it's negative negativity. So yeah. That's definitely tricky though, about risk-taking right.
[00:52:53] Ken: Yeah. You know, it's funny. I think that I consider myself a fairly conservative like risk-taker on social platform. But if I was trying to maximize reach and output, I would be significantly more controversial.
Like if you take a side on something that is almost certainly what drives interest and drives a lot of things. And unfortunately, As of yet, my goal is not to become like the biggest, most controversial data, data, YouTube, or broadcaster as it is. But I mean, that is an important thing. Like it, you know, if you're just giving like non spicy takes all the time, like everyone here is non spicy takes all design, right?
Like it's boring people. The excitement. People want to see like a battle in the comment section content section and like, you know, like a friendly one, but like that to me is, is really relevant. You know, something you also said about, about the haters on the internet is that there are, there are plenty of them out there and you try to listen to, you know, you try to take feedback when it is valid and, you know, something, something that I really struggle with is.
It's a silly thing, but my YouTube thumbnails, there's like one or two people that just keep saying that they're clickbait and they're going to unsubscribe and all of this type of stuff. And to me, like, that's hurtful. Like I work really hard. I'm trying to like, there's an an equation. Like you have to have some intrigue, you have to have some reason for people to click on the video.
To get people to watch for you to drive the most value to people. Right? If they don't watch your video, they can't learn from it. Right. and that's like part of the structure, it's part of the whole process of YouTube or LinkedIn, or any days you have to have a hook. Right. And I did a lot of, but to solve this, I just look at research.
I pull as poll questions. I pull the broader. Reach to my audience. And if, you know, like 10% of people are saying that my thumbnails were too clickbaity. It's probably too much. Right. But if it's a super vocal minority, whereas the majority really doesn't care and they find it fun and interesting. And they're like, Hey, this is funny.
Like you know, what kind of goofy looks in this? Like, using data, you can dispel all. You can really get down to what is truthful and not truthful. And rather than just like vote, letting vocal minorities really really crush your, your self esteem or your, your competence.
[00:55:38] Zach: Oh, yeah. It was like, yeah, you don't want the, like, you don't want to fall victim to like the squeakiest wheel gets the grease sort of mentality you actually got to look like, is there enough of these people here? Right, for sure.
[00:55:48] Ken: That, that has really all the questions that I had. I do have a couple user written in questions. We answered most of them actually on the. Throughout our, our general conversation one is, you know, they're very interested in when you're visiting Texas for a talk. I don't know if that's on your plan.
[00:56:09] Zach: That sounds interesting. Yeah. Texas is a good place and I love that state dude. Yeah.
[00:56:13] Ken: Well, I am actually contemplating a move to Texas, so I know the question is addressed to you, but to this person, you know, if you really want that, you can get the second best. And you know, David here is asking for what are some of the applications of data science or data engineering and more commercial products? Obviously, a bit more of a general question, but. You know, you have worked at quite a few commercial products here.
[00:56:45] Zach: I think there's a bunch, right? I mean, like you can use like just a couple that I've worked with, right. You can use data science to predict account compromise, right. And like cybersecurity. Right. You can use data science to predict like revenue and do revenue forecasting. Right. Of like how much how much revenue is the company gonna make. Right. That's like another kind of predictive sort of model.
And you can also do things like on the other side, like you can do data science to figure out like, Okay, which notification should I send to this person to have the highest chance that they clicked? And things like that as well. Those are, those are probably the three big ones that I've kind of worked on in my career. There's definitely the applications for data science are so vast. It's really astounding.
[00:57:29] Ken: Yeah. And they continue to grow and expand, which is sort of the beautiful thing. I mean, the funny thing with that question is like, the answer is kind of in. Right. The capabilities are endless and continuing, Oh, I guess by definition it infinite can continue to grow.
But there's I guess this question is probably more addressed me. It's how to eat a papaya in the most fulfilling way. I'll probably just make an entire video on that. So we'll, we'll skip that one for now. With that being said, where can people learn more about you? Are there any things you're working on right now that you want to share and you know, what's going on in your life?
[00:58:09] Zach: Yeah, for sure. So a couple of things like you know, you can find me in most places if you Google "eczachly". So that's E C Z A C H L Y. That's like my handle everywhere, you know, Twitter, Facebook, Instagram, LinkedIn. You'll find me everywhere with that. Another thing I'm on YouTube with automatic Data with Zack on YouTube.
That's my YouTube channel doing pretty well. I got a couple of videos there. I want to do a lot more like deeper kind of educational videos there. One of the things that. Like looking a lot more into recently has been like web three cryptocurrencies and like decentralized finance and stuff like that and that kind of space.
So I definitely want to do some kind of videos and stuff on that stuff. Like, so I'm going to be doing definitely a lot of stuff. They're going to be doing a lot more podcasts with people as well. So you're going to see a lot more of me, like. It's a pretty exciting.
[00:59:02] Ken: Oh yeah. Hell yeah. Well, I'll link all of those in the description and in the show notes as well.
So everyone should have a breeze finding you and learning more about you. Again, this was such an awesome conversation. I learned so much about your career, your experiences as a tech lead. A lot of information about data engineering as well specifically. So Zach, thank you so much again, and I can't wait to share this with everyone.
[00:59:27] Zach: Awesome. Yeah. Thank you for having me.