A 'Moneyball' Approach to Hiring Tech Talent

This talk with go into detail about what Neeraj looks for in a "Moneyball" algorithmic way, and that oftentimes it's good to focus on the unpredictable.


Neeraj Kashyap 0:00
It takes a lot of courage when you are outside of a community, or when you're outside, like you know, when you don't have that validation of your skills that credentials provide to actually step up and apply for a job. And so, giving those people a chance is generally a good policy to find misfits. And that's been very successful for me.

Lee Ngo 0:20
I love that.

Hello, everyone. Welcome to Educative Sessions, a podcast series with people in the developer world about their coding experiences. This is brought to you by Educative, which makes it easy for authors to provide interactive and adaptive courses for software developers. Today, my guest is Neeraj Kashyap, who is the CEO of Moonstream.io. And today --

Neeraj Kashyap 0:47
Sorry. Moonstream.to.

Lee Ngo 0:49
Oh, Moonstream.to! Excuse me, sorry about that! And we're going to be this is what I decided to title, our talk, which is focusing on the Misfits, and how to properly hire in tech, garage. Thank you so much for being part of

Neeraj Kashyap 1:03
this talk. Thanks. Thanks. Thanks for having me on.

Lee Ngo 1:06
Absolutely. So let's begin with, you know, you've had a very storied career in Silicon Valley. And yet your entrance into this community is a little unorthodox. So I'd love to let's as quickly as we can talk about all of your impressive work over the years in the company's been a part of and how you were able to navigate your way into these great companies.

Neeraj Kashyap 1:30
Sure, I started my life as a mathematician. And when I say mathematician, I mean, the type of mathematician that only works on pen and paper all day, every day. And I entered technology because I became very interested in healthcare. And I became very interested in diseases like Parkinson's disease, and other diseases that affect that, you know, degenerate the brain, like the policies and stuff like that. And, yeah, basically, after I got my PhD, I said, I just want to spend my life studying these diseases. So I moved to Japan, and I was doing research about these diseases for basically two years, I lived in Japan for two years doing that research. And you can't really do research on diseases like on pen and paper, you have to work with people, you have to work with devices, we were, we were coming up with diagnostic systems for those diseases, that could outperform a doctor. And that's how I got into technology. So we were using accelerometers and gyroscopes to basically analyze people's movements, to try and understand if they have dementia, or Parkinson's disease, or any other disease that affects the brain. Because these diseases -- diseases that affect the brain, they generally manifest themselves in terms of a person's gait, going off for, like, you know, people falling down, and things like that. So we were analyzing their gates in order to understand whether they had these diseases. And that's what got me into technology. The mathematical background certainly helped. But you know, that that's what made me fall in love with programming and with building software, and with building things that like, you know, other people use and improve other people's lives. Because when you're doing mathematics, when you're doing mathematical research, there's, if you're lucky, there's like five people in the world that care about what you're doing. And when you're writing software, like you know, you really know, every line of code that you write has the, has like, the potential to actually improve the lives of like a lot of people out in the world, which is just a different sort of pleasure that you get, you know, in -- from your work. So that's what got me into technology. It was pretty hard to make a living as an academic in Japan, and I didn't really want to do that, you know, for an extended period of time. So at some point, you know, once my postdoc in Japan ended, I decided, "Hey, look, you know, the place where I hear they're doing this stuff all the time is Silicon Valley. So let's move there and see what happens." So I moved to Silicon Valley. And initially, you know, the technology that I developed in Japan, I sort of made a business -- tried to make a business in Silicon Valley where, you know, we built that technology and product to retailers in the US. And that didn't work out. But that's how I ended up in Silicon Valley. And then, from there, you know, I mostly work in startups, I enjoy working with startups and in startups. I run the AI teams for companies like doc.ai, and, you know, Human API. I think they're good companies in healthcare, but I've also worked at Google. I've worked on the TensorFlow team at Google. And now I'm like, extremely involved in the blockchain. And I'm running my own company, Moonstream.

Lee Ngo 4:33
Wow. Well, that's incredible. And I'm largely amazed by how, like, you've transitioned to so many different things, which and yet you started out as a mathematician, right? It gives I think a lot of people potentially hope that you don't have to feel so fixed on just because you study the thing, therefore, that's the thing you do, especially if it's something that has a lot of potential applications. And what's it sounds like it's you've been largely motivated -- motivated by your own preferences and your own passions, especially for things like healthcare and AI. So, and -- I want to transition a little bit to what we're talking about today. So you've worked in all these different companies, and you've built all these teams. And you know, when we were preparing for this talk, it sounded like you had a radically different philosophy when it comes to hiring than what most people do. So, you know, some people look for certain degrees, institutions they come from, backgrounds, etc., etc.. But you follow a different algorithm, let's put it that way. Let's go more into that. So what are you exactly looking for in a candidate?

Neeraj Kashyap 5:40
You know, a lot, it's, it's, it's informed by my own experiences, you know, coming into tech, not not having been trained in like new technology beforehand. The important thing to realize, so what most people are looking for at least what I see on the market, as people are looking for credentials, right? So, if you're trying to build a team, and you're sort of operating at a certain scale, or beyond a certain scale, maybe you're not doing your own hiring, maybe you have like recruiters who are in charge of hiring or that kind of thing. And generally, what works well for candidates is that they have credentials, like, "Oh, do you have a computer science degree from Stanford, or from MIT or from Berkeley?" Or like, you know, any or some big school, like in what in your country? If you're hiring in a different country than the US? Or, you know, have you won some competitions, like programming competitions? Do you have like a really good score on LeetCode, or, you know, these are the kinds of things that people look for, they look for credentials. But ultimately, you know, I think, looking for credentials misses the point of what someone who build software actually does. It is a skill, but it's, it's also sort of, it's, it's also a craft, and people with different backgrounds sort of bring different things into, into this craft. So you know, just to - just to give you an example, I think like something like the best, some of the best programmers, I know, were not educated in computer science. They used to be philosophers, or they used to be historians, and then they -- and then they came in, they came into tech, because it makes sense to come into tech these days, then it made -- it's made sense to come into tech, like, you know, since the 90's. But they brought an analytical ability that isn't trained just by learning how to interact with the program. It's trained by doing deep analysis on any topic. I mean, the philosopher does deep analysis on philosophy, historians actually do deep enough, like deep analysis on historical events, right? Interpretation and things like that. The fact that they have that analytical ability, and they can use that ability to solve problems is really what matters. It doesn't matter whether they know like programming language X or like, you know, container technology Y. It really doesn't matter. What really matters is that they can solve problems - they understand how to solve problems fundamentally well, and then they can learn the skills that they need, like they can learn the programming languages and other technologies that they need in order to actually, like, bring that problem solving ability to bear on whatever they need to do as part of your team. And that's the important thing to understand, and that's what informs sort of like this different way of looking at how you bring people into a team.

Lee Ngo 8:22
Got it. Yeah. And I mean, personally, I appreciate that greatly. I, as like you, have not had a formal technical training in most things. And -- and I find that most people haven't actually, when you work in tech, and I think what you've done is, I think appropriately, demystified a lot of technology as requiring, like a specific skill set in order for you to thrive. Sure, that's absolutely the case. When you want to become, say, an engineer, or if you want to do a certain thing, you need to know this. But, as you I assume know and I've learned from my own experiences, technology is - it requires all the skill sets, it's -- you're, you're building companies, you're building communities, you're building ecosystems and environments, and you can't just do that strictly with engineers, right? You need to do a lot more. So, you know, let's say, like, you've got people with certain interesting backgrounds. Are there other things, like, beyond say fields of study or different styles of analysis that you try to look for when it comes to a candidate? You know, if there's a -- if a candidate doesn't have a lot of experience, specifically in tech, is that necessarily a deal breaker for you? I'd love to go a little bit more about the "misfits," and how, you know, we joked that you have a bit of a "Moneyball"-esque approach. I'd love to know a little bit more about what that means for you.

Neeraj Kashyap 9:44
Definitely. So, for the most part, what you need to look for -- well, definitely what I look for when I bring someone onto one of my teams is more sort of personality and culture. I don't you know, I'm not running like 100 people team or something like that. In general, no single person, I think should be running, you know, involved in the day-to-day activities of like, you know, more than maybe 10 people. So, the important thing is like, you know, what kind of a culture does your team have? That's something that you as someone who's building a team need to have a very clear idea about. For me, I'm looking for people who generally don't, who don't have any, like, you don't want to have like ego problems on a team. You want to have, like a sort of - I want to have a culture of modesty: people who are like willing to learn from each other, people who don't like, whose self esteem isn't based on like, how they're perceived by their direct peers, but like, you know, on their output into the world. I think that's like, much more important, And, who are willing to learn, and we're hungry to learn. Those are sort of the characteristics that I look for. And it's an interesting thing, you know, candidates who are more credential just on average, like, let's say that someone who doesn't have credentials, someone who doesn't have a computer science degree from a top university puts themselves into the candidate pool, then they generally have more of a growth mentality, because they were willing to take that first risk to even enter the pool of candidates. So they're there, they probably have like more of, on average, mentality to like, grow and to work hard, and to like really hustle, then somebody who's like, you know, who looks good on paper, because the people who look good on paper just have to work less hard to even sort of enter, enter that pool in the first place. So just as a numbers game, you know, if you if you really want to play "Moneyball," if you want to identify people who might be more of a cultural fit, look for those outliers that you know, where, like, you think at first, like, it's strange that this person applied for this job. Actually, that's where usually the value lies, because those are the people that are being skipped over for jobs by everyone else. But there's a reason that they're applying generally. It takes -- it takes a lot of courage, when you are outside of a community or when you're outside, like, you know, when you don't have that validation of your skills that credentials provide to actually step up and apply for a job. And so giving those people a chance is generally a good policy to find misfits. And that's been very successful for me.

Lee Ngo 12:17
I love that! Because I feel like many of people who are looking to get in are looking for that opportunity of inclusion know that experience of just not being right because they just didn't decide that one thing, but I like that you're acknowledging that and hope. well -- not too much because it then it messes with your algorithm. That's the thing is like, if everybody catches on to this mentality that it's just like what happened with like SABRmetrics, like, everyone started doing it, and then it started to become less and less effective. But I think this is at least a good thing for a lot of people who are interested or inspired into technology that, you know, it's not too late, in that there are absolutely at the top echelons places that are not going to penalize you for having a different background. In fact, if anything, they're encouraged to look at it. So that's, that's great to hear. So we've come to the end of our questions, Neeraj. I know this goes by really fast, I wanted to give you an opportunity to talk about Moonstream and or you can talk about whatever you like. Really, the floor is yours. Go for it.

Neeraj Kashyap 13:21
I'd love to talk about Moonstream! So, you know, what we're building at Moonstream is an open source platform for blockchain analytics. So if you know, if you own crypto or if you -- if you develop for blockchains, like Ethereum blockchain or something like that, it's generally pretty difficult to understand what people are doing on the blockchain. It used to be easier, you know. The community used to be smaller, there used to be fewer people that were actually doing stuff on blockchains. And at that time, like, you know, these communities functioned as villages where like, you know, you could like, just sort of follow along with the local gossip, to know, like, what was going on. But now there's, like, you know, 10 million people participating in blockchains. More than 10s of millions of people participating in, in the various blockchains, and it's no longer a village, right? It's a big city. So how do you know, like, what's the important stuff going on, on these different blockchains that that's pertinent to you, that could be a value to you? And you can't do that just by like, you know, listening to the local gossip, because like the the Discord server that you're on probably doesn't know about, like this huge activity that's gonna happen over here in the next couple of days. And there's no real way to find out about it before it happens, unless you have, you know, you have machine learning algorithms that are processing the data, the public data that's available on all these blockchains and surfacing to you the stuff that, you know, that could be relevant to you and could be important to you. And that's what we're building with Moonstream is, you know, we analyze all of the data that people produce on blockchains. We understand like, using static analysis, like even like what's going on with the programs, what are the semantics of those programs. And then from that we understand, like, "Okay, this is relevant for this type of activity," and so if you sign up for an account on Moonstream, you can say "I want to know like, you know, whenever there's a new NFT that has been minted but hasn't been like released yet, because then I want to sort of get in on that action." You can you can sort of set up a trigger at that lower level of abstraction and get alerted when we find activities like that on the blockchain. So that's, that's Moonstream. We have a community on GitHub. So if you just search for Moonstream on GitHub, you'll be able to find us. We also have a community on Discord. You're welcome to join us. That's, that's us.

Lee Ngo 15:31
Fantastic! And it sounds like a lot of what you're doing now is starting at the community level, but I think that's, I mean, I'm a community guy. So like, that's the way to go about things. Neeraj Kashyap, thank you so much for being on our show. And thank you for contributing your fantastic wisdom and your knowledge on such a vital subject for a lot of people. And I want to thank everyone else for either listening or watching this podcast as well. You can look at more episodes on our YouTube channel, or you can look us up on pretty much any podcasting channel at Educative Sessions. And lastly, if you want to learn a little bit more about Educative, check us out at educative.io. So for all of us here at Educative, thank you so much, and happy learning. Bye bye now.

Hope you enjoyed that session. This episode is available on YouTube and also on many podcast platforms. If you'd like to be part of Educative Sessions The form is open now to apply via the link below. You can also email me at lee@educative.io. Lastly, don't forget to like and comment on our content. Be sure to subscribe for us as well. And of course you can learn more about us at educative.io. Happy learning!

Similar posts