What is the best career path for a new developer? Should you go to college? What about grad school? Regardless of your path, getting involved in the open-source community is a smart way to learn, collaborate, and build a strong portfolio.
In this interview, Senior Software Engineer at Google Research Sandile Keswa discusses hybrid learning models in education, how to get involved in open source, and advice for up-and-coming developers.
Episode Resources
Subscribe
Transcript
Jon Gottfried: Hi everyone. I’m Jon Gottfried, co-founder of Major League Hacking. I’m so excited today we are talking to Sandile Keswa who is a Senior Software Engineer at Google. Hello.
Sandile Keswa: Hello. How are you doing?
Jon: Pretty good, man. It’s good to see you again. Sandile is a front-end leaning software engineer. He’s a huge basketball nerd, great sense of humor, loves coffee, which I think he’s drinking right now, and puzzles. He loves hacking on consumer-facing side projects that augment the human experience, which we’ll talk about a little later. Sandile is actually an MLH alum. He ran a number of MLH events and helped build the hacker community at Temple University which is where he got his BS in Computer Science and Mathematics a couple of years ago. I am so excited to have you here, Sandile.
Sandile: Happy to be here. I don’t do podcasts often, this is new.
Jon: You’re more of a listener, not a speaker?
Sandile: Exactly. I’m what you call a lurker, I guess on Reddit.
Sandile: I’m here for the content, but I never make any of the content myself. Now we’re on the other side.
Jon: Here you go, man. This is your start.
Sandile: Oh my goodness. I’m breaking out. Great.
Jon: Awesome, man. I want to kind of start where a lot of folks in our community are right now, which is beginning their careers. I know you’ve been at Google, I think about five years now, right?
Sandile: Yes.
Jon: You went through that career path of an individual contributor like software engineer, did amazing work, became a Senior Software Engineer. I’m sure you’ve worked on some pretty amazing things over the years, I’d love to hear more about that. What are some of the projects you’ve gotten to work on over at Google?
Sandile: Totally. I started working at Google like you said, five years ago or just about, I don’t know, give or take a few months. It was right out of school. The first thing I worked on was in advertising which at the time I was not thrilled by. You don’t associate advertising with interesting problems to be solved because from the user side, you just go to a website and it’s just blocks of stuff, it’s like billboards right in your face. Working in ads really showed me exactly how vital the ads ecosystem is to what enables the web to be as magical and interesting and wide-reaching as it is now.
Display advertising especially really unlocks a lot of those key features that you love about the web. You can go and read somebody’s blog and that blog might be able to be self-sustained just because you get a little bit of ad revenue every time an impression comes through. That is pretty cool when you think [crosstalk] you don’t need money to enjoy some of the beautiful and interesting content that we have on the live web.
I worked on a team called display ad verification which sounds like the most crufty corporate-y– It’s not a very interesting sounding problem, but it is actually quite interesting. This team was specifically dedicated to making sure that– This is the example I like to use, is like, airline companies don’t want to put their display ads on articles about plane crashes. That’s not.
Jon: That would be very counterproductive.
Sandile: Exactly. Disney doesn’t want to put their ads for the new Mulan or whatever on pages about–
Jon: Car crashing or something. Like terrible news.
Sandile: The interesting thing about that problem is people don’t actually really ahead of time always know, but they don’t want to be placed next to in the real world. This team is basically dedicated to figuring that problem out. Giving advertisers the tools that they need to make sure that their ads appear in the right spots. DDM verification taught me a lot of stuff mostly about how machine learning is very hard. It’s really difficult to get it exactly right but when you do, you can enable some incredible work.
That was DDM verification. I worked there for three years. Then after that I transitioned to Google Research at New York, focused on education. Education’s always been a passion of mine. I don’t know, it feels almost like the thing that unblocks all avenues. Education is pretty much at the root of any interesting thing that any person might want to do. You always need to find out more. I was working on a few education-related research projects.
One of the research projects that I was working on ended up kind of graduating so to speak and becoming a thing that we’re actually investigating actively and putting resources into. That’s under the Google Education Umbrella. Google Education in this past year is at a very exciting year, I guess
Jon: Transitional year maybe.
Sandile: Some products that you might have heard of, Google Classroom, Forms, ChromeOS, products like that. It has been such a huge learning experience with all the changes that have happened with social distancing and learning from home and things like that. It’s been a very interesting time to be in the space that’s the intersection of technology and education. Lots of really interesting problems to solve and a lot of problems that have never had to be solved before. How do you even know that what you’re teaching somebody is sticking if they’re just a picture on your screen that moves from time to time? It’s really difficult to evaluate those sort of things.
I feel like a lot of the stuff that is really important to the teaching process is soft stuff. Stuff that doesn’t show up in the numbers like oh, this student has been daydreaming a lot lately. I wonder what’s going on. I’m going to go and dig in, going to go have a conversation. It’s hard to pick that stuff up when we’re behind these video streams, stuff like that.
Jon: Are you saying that technology can assist with that?
Sandile: Yes. There are a lot of ways that technology can be adapted to be more helpful and more contextually aware of specific things that should or should not be happening I think in education. Broadly speaking, the video call was a basis, a foundation for making a lot of these use cases possible. It’s not smart enough in its current form to really facilitate productive classroom learning. I think a good example is keeping younger kids engaged. How do you stick a six-year-old in front of a computer for six hours and expect them to stay engaged the whole time? That’s crazy.
There needs to be an accounting and some innovation in that space to really figure it out. Ultimately I think some aspects of what we’ve taken away for 2020 are going to stick around, it’s going to make more and more sense to enable some of these cases where students can stay home and still get a great educational experience without being, in close proximity of their teacher.
Jon: It definitely increases accessibility in some ways. I think even for us, right at MLH, we were historically this totally in-person community. Now we’re really thinking about it as a hybrid model because it allows a certain amount of creativity that’s not bounded by physical, space, and requirements.
Sandile: Absolutely. I think we’ve all been challenged to think about work and school in different ways the past year. I don’t know, you even look at the technology industry and how the advent of remote working posts have really skyrocketed. I wonder what that’s going to do to the concept of that split between work and home and how we manage a virtual workplace.
Jon: What is a work day, right?
Sandile: Yes what’s a workday? Absolutely. Some of these notions that are as old as the industrial revolution we’re kind of taking them apart and doing a reaccounting of what’s valuable and what isn’t. That’s pretty exciting from the perspective of I wonder what that will enable down the road.
Jon: It’s funny though that you mentioned that getting a six-year-old to stare at a screen for however many hours, six hours, eight hours, is incredibly difficult. Every six-year-old I know could play Minecraft for six hours. What is the difference between that and what they’re getting in the classroom?
Sandile: I think there is a Twitter thread that I was reading a couple of days ago. It was talking about some comments that Elon Musk made about I think we’re overthinking some aspects of education. I was like, Oh okay that’s a cool thing to say, but I think his point was basically that. You don’t have to ask a kid to play Minecraft they’re just going to do it and they’re going to do it forever. It’s limitless boundless energy for them to do that. It’s like, can we channel that same avenue of energy and curiosity, and passion towards a direction that has educational outcomes? 100% we can, we just haven’t gotten creative enough to really make that possible yet.
Jon: I really distinctly remember, I grew up in the final years of dial-up internet when computers were actually a pain to use. I remember having these arguments with my parents where they’d be like, “Hey, you’re on the computer too much, go do other stuff, blah, blah, blah.” Precocious little six-year-old me would be like, “No, no, I’m learning stuff.” The funny thing is, I bet you kids who play Minecraft are saying the same thing. I bet you they’re modding and figuring out the circuits with whatever those red bricks are.
There’s an argument to be made that education should be more fun and creative like that and I think hackathons actually represent a lot of that creativity. You did a ton of hackathons, you organized hackathons but you also got a CS and math degree, which is highly academic. How did those different things contribute to who you are as a developer?
Sandile: Right. Yes, you know what? That’s a fantastic question. I don’t know. If we run it all the way back to undergrad, and in the hackathon scene.
Jon: The olden days.
Sandile: The olden days, if you will, yes. I think there is value to academia. Dedicating yourself to figuring out some really difficult methods of thinking and learning some history behind some of the important things that drive progress today. I had a realization in undergrad, and it was that I think being a software engineer in the modern context is more being a craftsman than being a scientist. It’s more about the techniques, and the tools and the methodologies, and the culture of it, than it is about understanding necessarily the minutiae of the theories that contribute to cool things like databases and tc, right?
Those things are all really important to understand theoretically, but you don’t need to understand those things to build an application that helps people get in a taxi and get to the airport. You don’t need to know that information. Computer science, I think it’s really interesting cross-section between practicum, basically the idea of engineering and those sets of skills, and theoretical exploration and figuring out new discrete mathematics to solve high-level problems.
Distributed systems are a good example of where computer science is necessary to building things that are actually practically helpful. You need consensus algorithms, you need all these systems dedicated to distributed ledger management. There are all sorts of things that you have to figure out that are very down the scientific route and avenue, but then you bring them over, you implement them in the form of a horizontally scalable MySQL, now, it’s practical. Now it’s a tool that people can use to build great stuff, so I don’t know.
When people ask me if you need to go and get a degree, it really, really depends what you want to do. If you want to build the applications and services that people will use and need and want, I would argue that you don’t need a computer science degree to do that. If you want to build that infrastructure, if you want to advance the state of the art, unfortunately, you’re going to have to hop in. You’re going to have to get theoretically sound and really figure out the innards of discrete mathematics and computer science and all that good stuff.
Jon: Yes. It’s like if you were to compare the people who do molecular gastronomy with a really good chef. The chef may understand how to apply their skills incredibly well but they’re not looking at the chemical reactions that derive a certain outcome. I think computer science is somewhat similar, where you’re not going to be implementing a Bubble Sort algorithm, you’re going to use one that’s in a library. Or probably a better one than Bubble Sort, anyway.
Sandile: Exactly. I know it’s a completely different skill set to write a new Bubble Sort or use an existing Bubble Sort and decide how to build maintainable software. It’s a completely different set of concerns. Whichever one of those things you want to do should really color how you structure your career and structure your time.
Jon: Yes. Google, famously has a lot of these foundational technologies available as tools, and as utilities for other applications. MapReduce is probably the most well-known example, but I’m sure there’s many others that someone did this crazy research project and came up with an incredible solution that’s now used by thousands of developers or millions of developers.
Sandile: Right, absolutely. That part of Google’s ethos is one of the reasons why I wanted to work here in the first place. I feel like there are very few large companies where I could go from display advertising to educational research with minimal friction,Like. Yeah. Just like switch, you know, and. These are fields that are usually completely unrelated, but at Google, I went from–
Jon: It could be totally different companies.
Sandile: Yes, absolutely. It’s like at Google, I went from the fourth floor in New York to the sixth floor. That’s what changed, you know what I mean?
Jon: A big change.
Sandile: I’m a fan of that but I think it’s given me time to reflect. I feel like as an industry we need to advance culturally towards dedicating ourselves to actually solving problems. There’s this weird split, I feel like in technology culture, where everybody says that they’re making the world a better place and it’s like a meme because they’re building developer metric tools. It’s like, “Yes, sure, the world is slightly better now than it was before this metric tool.”
At the end of the day, we’re at this super interesting time in history, where we have a bunch of big complicated very intimidating problems to solve and we’ve never had more tools and more incredible people working on those problems. If we incrementally build up to focusing on those problems more and more, even if it doesn’t have necessarily an economic benefit. I think we should, and I think Google’s really blazing the trail in that category.
Jon: Yes. I know you’ve done a lot of work on your own in that way through the open-source community as well. I think in many ways, that’s the perfect representation of democratizing really complex technology in a way that benefits a lot of people regardless of the business value, and you can make an argument about the business value being there, but somewhat disconnected. There’s nothing else like open source in any other industry. It’s just really novel thing. How did you get involved with that? Because I think for a lot of developers, they go straight into the corporate side, and everything they write for the rest of their career is proprietary.
Sandile: Right. My involvement with open source goes back to when I was in undergrad and sitting around my dorm room figuring out what to do. I would say that hackathons helped me find a friend group of that type of person who likes to create things and tweak things and mess with stuff, figuring out how stuff works. The natural next step to having that sort of group of people with you is to build stuff that other people can use too. It’s this realization that like, oh, the hacking that you just did on a Saturday night, oh, you figure it out how to get like, when you press this button how to make a bunch of different LEDs start blinking onThat’s terrible information.
That was my beginning into open source is building a bunch of really stupid little projects. Every once in a while one catches on a little bit, and you meet some people that are really interested in it and you get some pull requests. It gets you even more invested in that community. Open source is one of the coolest things ever. If you just step back and think about what it is, it’s a bunch of people doing work towards things that are necessarily beneficial to somebody, somewhere for free, which is awesome. It’s very cool. It’s a lot.
You can find some of the most incredible engineers and thinkers and creators on the open source playing field, so to speak. You go on GitHub and you can learn so much about how to build great software and how to reason about difficult problems just by interacting with and reading people’s code. Which I feel is such a competitive advantage in the technology field. Where it’s like you can literally educate yourself. Get to a level of proficiency that I don’t think in any other trade you could do through the same needs, GitLab and all of these other social constructs built around open source are really powerful and they have network effects that make us all build better and greater and more interesting things
Jon: I was obsessed with SourceForge growing up, which was the precursor to GitHub in many ways. You’re right that it gives people this ability to gain proficiency, but it also gives people an ability to have an impact. How else could someone who doesn’t have a job, but has skills and passion, create something that millions of people use? Obviously, you have to be able to pay the bills, we can’t discount that. That’s interesting and many people who do open source have a full-time job that’s not open source related, but like any student could go on and make a pull request that goes into a project that’s used by every company in the world. That’s wild. That’s never existed before in the history of humankind.
Sandile: That collaborative aspect to solving complicated problems is so encouraging. That maybe these magnificent, huge issues that we all have to solve are indeed solvable If we all contribute a little bit towards a solution, that’s a really cool idea.
Jon: I feel you’re touching on two parallel trains of thought here. The technical problems that need solving, but then there’s also the practical user facing problems that need solving. I think you said on your website that you build things that augment the human experience. What does that mean, exactly? How do you balance those two competing forces?
Sandile: I guess what I figured out early on is that I want to build things that make a real tangible impact to human lives. I want to build things that make it easier to learn how to fix your tractor. I want to make stuff that makes it easier for you to figure out when your crops are going to need watering so that more yields, right? Those are real, tangible, actual problems to be solved that directly affect, the things that humans need to do to make this society possible. There are lots and lots of problems that technology can solve. It can be very distracting to go and solve all of these immediate problems that don’t necessarily immediately impact what it is to be a person.
I think augmenting the human condition is my way of saying I want to make things that make it easier for people to be people. Whatever that means. That’s a really lofty thing to say, but it can be things as basic as making it easier to call your grandmother or making it easier for you to keep track of her recipes so that when you make that shakshouka, that smell just brings you back to childhood.
That stuff is so important to what it is to be human, in my opinion. Technology is so powerful in those specific use cases. It just closes gaps. That’s the stuff I really aspire to build. I acknowledge that those are some of the hardest things to build. There are all sorts of ways to mess that up. Those are really difficult problems to solve. I think we’re seeing in the wake of the last few years of technology news that bringing people closer together isn’t always a silver bullet. It’s a complicated set of issues that you have to reconcile before you can really create a beneficial set of network effects thereafter.
Jon: I also feel what you’re saying, augmenting the human experience your definition of it is very I guess I would say personal. It’s about how do you improve the things that people do to connect to one another. If I had to rephrase it. I think there’s a whole movement of augmenting the human experience towards increased productivity, which is a twist on what you’re saying. It gets down to a really philosophical argument of how does being productive impact the human experience, and what are the things that someone should be optimizing their life for, and how does technology enable or in many ways build that outcome?
Sandile: Productivity. That’s a good question to ponder on and meditate on. What is productivity in the context of what it means to be a person? I don’t know. I feel so in the economic conditions that many of us live in, Productivity is a very necessary thing to optimize for, because it directly affects what you can, the helpfulness that you can offer to people in your life and your family, how you contribute to society.
Jon: The value of your time.
Sandile: Exactly. Technology is, I guess it’s superpower is automation. It’s really good at taking something that’s very repeatable and then shrinking it down and saving you all the time that you no longer need to spend doing that re repetitive thing. I would say building productivity tools and automation-type tools are very definitely like super duper important. As it pertains to solving those big problems. Productivity software. Definitely, it makes a mark, it helps us move the buck forward.
I don’t know. I feel in the society we’ve constructed, we really overvalue productivity from an economic sense. It’s just capitalism, really, it makes you really want to produce more and more and more, and consume more and more and more. I think that’s only a certain subset of problems that are optimized by increasing the efficiency of human beings. I don’t know. That’s a pretty sketch.
Jon: It’s a big question.
Sandile: Yes. I have no idea.
Jon: I think that somewhat menial required tasks make life more interesting in some ways. I enjoy going grocery shopping. Sometimes it’s a pain in the butt. I don’t really want to do it in the moment, but the fact that I do it. I think is actually valuable to me. Even though I could trade that time for money to have someone else do it for me. I always reflect on that where it’s, do I want to devote more hours to let’s say work for example or something else or do I want to just do the required things to live a life and I don’t know. I think there’s a balance there and there are many things that I pay to not have to do. There are many things that I do because I think that even boring stuff has value.
Sandile: Absolutely. I totally agree with that. I was talking to a friend of mine actually a couple of days ago about this thing that he was trying to do with machine learning. Basically [chuckles] it was handwriting analysis right, so a bunch of people were writing math expressions and we’re trying to turn those Like turn those.
Jon: There was a winning hack at PennApps that did this probably not quite as well.
Sandile: No. It’s a really it’s interesting problem to solve and it’s directly helpful, being able to intuit about and reason about mathematical expressions that somebody’s writing on a piece of paper, that’s something that needs to be done. He was having a bunch of issues with all of the different ways that people could draw like the F in F of X right. It’s like all sorts of different notation styles and how old the person is that’s running it has a huge impact on just what it looks like, how reliable it is, how consistent it is. Every time you write that expression there’s a lot of variance.
The model was failing he couldn’t get the confidence that it roll up. At a certain point, he had like a plateau and he was hanging out at home and he had a bunch of samples that were really inscrutable, really, really hard to read. His girlfriend peeked over and was able to with 100% confidence just go through and identify each of the different F of Xs and identify what all the expressions are. It’s just this strange realization that, I don’t know, sometimes it feels like we all have the most powerful computer ever created in the history of the universe in our heads. I don’t think we really found a great way to combine that with technology, we’re trying to usurp it with technology.
If you look at self-driving cars same example right, it’s like we’re trying to get Level 4 self-driving which basically means we’re just trying to get a computer to do all the things involved in driving. If there’s a bouncing tire on the road, the computer needs to reason about that. If there’s somebody that wants to cross in front of the car the computer needs to reason about that. If somebody’s trying to merge in front of you, the computer needs to reason about that.
That is an incredibly difficult problem to solve because first you need to solve cognition right. You first you need to be able to have a computer that can reason about a world arbitrarily without any context, that’s something that we haven’t figured out how to do.
Jon: That’s something the brain is incredibly good at.
Sandile: Exactly right. If we could find a way to mix those two, use the strengths of both types of machine, human and robot I think that’s really the path that we have to forge better than we currently have. There’s a lot of I guess human-computer interaction research to be done in that particular avenue and I’m excited to see where we go with it. I think the technology industry is over-obsessing with automating everything, not everything is most optimally automated by a machine.
I don’t know how we got onto this topic, but–
Jon: It’s a good one. Are you a big sci-fi fan?
Sandile: Yes I am a big sci-fi fan absolutely.
Jon: I feel like what you’re talking about it’s like Asimov, right? It’s like iRobot, Caves of Steel, those are some of my favorite books as a kid.
Sandile: Sci-fi is a fantastic looking glass for human beings, it’s a great way to stand back and look at where this trajectory goes and all sorts of different options. It’s like some of my favorite books in the world. Right now I’m reading The Three-Body Problem.
Jon: I read that recently.
Sandile: Which is a great rate highly recommended, but there are a bunch of really, really interesting takes on all these different scenarios that could affect where humans go next and sci-fi is really great for that
Jon: It shows the best and worst-case scenarios of how humans react to and use technology.
Sandile: Exactly and to some degree like humans in technology now that’s a relationship that goes back for a very long time. The thing that made our species different really is being able to use tools or technology to get advantages from a fit standpoint. Tools are what make us human, technology is what makes us human in a certain way which is a weird thing to think about. I think it’s really important that our relationship and our usage of technology is done in such a way that it enhances what it means to be a person and meditating on what that means is really important.
Jon: I’m curious are either of your parents developers?
Sandile: My dad was a mathematician, financial stuff, but is now a data architect. I forget how that happened. [laughs] I guess yes. Then my mom was a geologist but is now a data scientist so I don’t know how that happened either. Technology I guess ate the world overnight.
Jon: Well all things converge on data science maybe?
Sandile: Basically. It seems to be like a really powerful force outside of the technology sphere or I guess the technology sphere is just maybe growing to encompass these things so they both oddly enough write code now.
Jon: I was going to ask how they would explain what you do, because my parents have never been to a hackathon and don’t get it at all and they probably do get it really well.
Sandile: To a certain extent they do, but in some ways they don’t, I think especially for my mom, technology it’s just like a means to an end. It’s like yes, I want to analyze this set of data, I will use pandas and I will use and I will use all of these, I’ll use carrots to analyze the information that I have and create a model and whatever. Then I’m done I did it, I analyzed the data. Great. Moving on.
Whereas I think being a hacker is a much more intimate relationship that you have with technology. Technology is like your collaborator while you do cool stuff. Sometimes you just do stuff to do stuff. Explaining what it means to go to a hackathon, I remember my dad was just like, “Why are you doing this over and over and over again?” [laughs] It’s a great question, but it’s really because being a hacker is being a creator, it’s self-expression at the end of the day. It doesn’t I’m sure feel like it or look like it to first clients, but that’s really what going to hackathons is, using technology as your paintbrush and your canvas and painting something nice.
Jon: I love that, I always think of hacking as this intersection of art and science and we were doing a bunch of research years ago to figure out how do you describe the ideal of a hacker right? What words can you use? We happened upon this term “autotelic”, which means someone who gets pleasure from the process of something rather than the outcome.
I feel like that describes a hacker really well because even when a project fails often people leave feeling like they had this incredible learning experience or social experience or just tackle something that was creative and difficult and the output is not necessarily the point.
Sandile: Some of my fondest memories are failed hacks. [laughs] It’s like hacks that didn’t go anywhere. I think it was like PennApps 2014 fall or spring I forget which. Me and a friend of mine were trying to build an audio social network so it’s something like approximating– What’s that?
Jon: Clubhouse, is that it?
Sandile: There it’s, it was very Clubhouse-y.
Jon: You were a couple of years too early there.
Sandile: It was a little bit early, I missed the window, but yes, we had all sorts of issues with getting the recordings just right and our sharing model and all of that stuff. What I remember most is that at Temple, we would bring basically all of the people that would hack around on weekends and stuff in the ACM office. We would all just descend on a hackathon and we did that all the time and a lot of people that we brought with us just started trolling our application. This is like they hopped in and they started– This is like this really weird thing.
This is what happens when you get delirious. It’s like 4:30 AM, up on like a Saturday night and you’ve been awake for 36 hours. This is like people hop in, and they would just fill up our database with mouth sounds. They just eat really close to a microphone, it was just the most–
Jon: ASMR.
Sandile: Right. They put great comments on it, you think they’re going to say something really cool. It’s like, “Man, oh. I recorded a song.” “A song? Oh, great.” Hit play and it’s just like do like this, just somebody eating something disgusting like yogurt. It’s visceral mouth sounds.
Jon: There’s got to be a Clubhouse room that’s just dedicated to that.
Sandile: Yes. That’s one of my favorite memories. It was the most ridiculous thing. We ended up not even submitting the hack because it was so incomplete, but it was so much fun. I laughed for– It was great. That’s absolutely– I completely agree. Going in hacking and doing something like that is so much about the process, and so little about the product.
Jon: How similar is that to something like Google Research? I would guess those are very moonshotty products. What was that like?
Sandile: Google Research is definitely a little bit more structured I would say. Most of the projects that are ongoing are based on a particular paper or concept that’s explored by a bunch of subject matter experts. Eventually they pull in some engineering support, pull in some project managers, and it might graduate to be something more interesting. I guess I haven’t thought about it that way, but it is a very similar situation to the hacking idea where you’re just meditating on a concept and really seeing where it takes you because you don’t know.
Oftentimes, when you go to a hackathon, you start with one idea and you finish with another one. In research, I think it’s very much the same way. You start thinking that reality is some way, and then you find out that it’s another one. The product that results from that or doesn’t result from that is very much a- it’s fluid, and it really depends on those. Google Research is very much like that. Things get created and fade away all the time. You’re just trying to find something compelling enough to actually make it out there into the world.
Jon: Interesting. I would imagine that as you’re working on these more bleeding edge products, that they’re– You mentioned that you’re meditating on an idea. and working it. How much of your time is spent doing versus experimenting and learning? How would you actually break that out?
Sandile: I would say lately, it’s been a lot more doing, because the early phases were more research-y. More like, I would build a prototype that looks like this, we’re going to do a few UX trials, see what happens. Does it affect learning outcomes? Maybe, maybe not. Maybe let’s try it on 8th graders instead of 3rd graders. It’s much more iterative. We were tracing the shape of something interesting, though we finally found the heart of it I think a year and a half ago, and now we’ve just been building it actively.
Jon: What is it, if you don’t mind me asking?
Sandile: Well, it’s not released yet, so I can’t-
Jon: It’s a secret?
Sandile: I can’t talk about it in-depth. Basically, it’s a next generation educational learning product. Very much.
Jon: Very mysterious.
Sandile: Yes, very mysterious. I know. [laughs] It sounds so ridiculous to say, “Oh, great. Just wait.” We’re still very much in the weeds of trying to figure out how it’s going to work and who it’s going to be for.
Jon: It’s going to be like The Matrix, right? Where you just go into the simulation and you snap, and you know a new skill.
Sandile: That’s exactly what it is. Perfect. It’s going to change the whole game.
Jon: I got inside from that info.
Sandile: Probably won’t be like that, but it’s– Yes, we’re definitely seriously going back to first principles on what the learning process looks like in the classroom and how you literally get from the base floor of a concept to the top floor. It’s a bunch of journeys that like everybody takes and, generally, I think in education we kind. Summarize for everybody, right? We generalize the process that everybody takes. “Oh, you learned long division.” Then after that, somewhere along the lines we graduate to algebra, and once you know algebra, we have relations. We’re going to take you to “Do you know the quadratic formula?” It’s like this path that is so well trodded that you almost have to take it.
In these different interesting and sometimes really challenging areas of education, I think it makes sense to take alternate routes and build tools that facilitate that exploration.
Um, and that’s sort of what this product is, is somewhat about. It’s um, it’s about
making those tiny journeys like a much more formalized part of the educational process. Make it a lot easier to learn things in alternate ways. That’s the idea, but as you can tell, its very lofty and weird. You can do that. That’s a great question. We don’t know we’re figuring it out. Maybe we’ll never figure it out, but that’s the beauty of research. You’re chasing after the golden snitch, and making me catch it, and maybe you don’t.
That’s what I’m up to right now. I’m mostly we’re in the building phase of that. Basically we have a bunch of different ideas about what we want to actually come up with. and now were just trying to build up all of the foundation of the how, so to speak, then see what happens next.
Jon: It’s interesting when you’re talking about education. I’m reflecting back to the ways and places that I have learned the most. I think for me, a lot of that was heavily dependent on the person who is teaching me it. Great teachers, great managers, great mentors. Have you had people in your life who were so incredible that they influenced how you think about education?
Sandile: I guess my number one mentor would be my dad. It’s like a–
Jon: Aww.
Sandile: Maybe a cliche answer. My dad was raised in Apartheid South Africa, and education is really how he escaped his situation. It really took a few special teachers thast enabled him to come to college in the US and build a whole life and a career, which when he was 14, he had no conception of that’s what happened. That’s the power that education brings, it broadens your world completely. It’s also the power of great teachers that do that as well.
In raising me, I think, he instilled in me this reverence for the importance of education and the importance of the soulful atmosphere build in educational contexts like classrooms and student teacher relationships. I think that’s really guided what I do for a living. I think that stuff is really, really important and really, really difficult to do.
From a broader mentorship angle, there are a bunch of programmers and thinkers and creators that I think really influenced me and continued to influence me while I try and build the real things. you know, like Ralph Levion is one of my favorite product engineers. To read his musings on how to build front ends. And more better and more performant and more interesting ones. Iheanyi Ekechukwu is doing some really great stuff in the front space as well. I follow him.
Then on the other side of things, Bryan Cantrell has made some of my favorite talks ever and he has all sorts of resources on how to be a great manager and TL. Brian Miles, same thing, a fantastic technology leader. Finding people to model your game after so to speak, is really, really helpful. Especially if you ever get to have a conversation with one those people, you can learn a lot.
Jon: Yes, absolutely. That’s awesome. Is there any advice that you might offer to up-and-coming engineers who are maybe where you were five years ago?
Sandile: Yes, totally. I think there’s a lot of information out there, like recommending that people go and do boot camps and read all sorts of resources and go and take specific classes. I guess this is biasing towards my own experience but the best way to learn how to build a chair is to build a chair badly. That’s just how it works. If you want to learn how to build software, build software poorly.
Do it, go out there and build the most ridiculous, useless barely works piece of something. Then throw it at your friends and see what happens and then continue iterating on that, keep building. Because ultimately, it’s like I was saying before, is like software is a craft, it’s a trade. It’s not really so much of science as people expect it to be. The best way to learn how to do something that is to apprentice under somebody or work with a bunch of people who know how to do better and you’ll learn so much so quickly.
That’s my recommendation. I guess that recommendation doesn’t really address the should I go to college, should I go to grad school? Where should I work? That core mantra of continue to build and build better every time has really instructed my path, I think, through the industry.
Jon: I mean, be prolific and people won’t remember the stuff in the beginning, right?
Sandile: Yes, pretty much. Pretty much. Everybody conveniently forgets all the failures until the first success, I think. Those failures are so important, which of course, is a cliche thing to say, but it’s so true. You really can’t succeed without failing a million times first.
Jon: That’s good advice. I would like to end on something that’s totally unrelated technology because I think that everyone I know who is a good programmer has a lot of varied interests outside of tech. What’s something that you’re super passionate about these days that’s not tech related?
Sandile: I think it was during the pandemic, I’ll do multiple poppy phases.
Jon: Well, you’re like many people. I still have my sourdough starter.
Sandile: Oh, yes. The sourdough starter I guess it’s like a common shared experience at this point, everybody’s got one.
Jon: The memes.
Sandile: I think the thing got me stuck, or I got into woodworking. Just building stools. Well, I built a stool and I built some picture frames. I was like, “You know what, this is a pretty dope.” I think the thing I figured out is that it scratches very much the same itch. Is like as hacking or programming and stuff that you get something physical at the end of it.
It’s like, “Oh, my gosh, this is great.” It’s a lot of fun and allows you to I don’t really sell and out and really figure stuff out because it feels like something that should be obvious and easy. Then you try and do it and it’s like, “Wow, it’s actually really hard to build something great.” It feels so analogous software but different, but the same, but different.
Jon: Are you impacted by the wood shortage?
Sandile: Yes, big time. I’ve been on craigslist with my binoculars looking for– I almost got a hold of some walnut that somebody in the South Bronx is trying to get rid of which would have been great because walnut costs a fortune right now. You got to get creative nowadays to try and get wood. [laughs] It’s difficult. It’s difficult.
Jon: I had the same thing with flour, man. When I got my sourdough starter, every grocery store was out of flour and I ended up having yet a 50-pound bag from some restaurant supplier that I’m still going through.
Sandile: [laughs] You might be going through that for a bit. Absolutely.
Jon: Yes. It’s a lot.
Sandile: Oh, yes. Oh, yes. The other thing that I think has been taking up my time lately is just somehow I’m a basketball nerd. The NBA added this thing, which has changed the whole game, which is called the play-in tournament. It’s like, normally how the NBA Playoffs work. It’s like you play the regular season and then they take the top eight teams by record in the East and in the West and then they do a series based on that.
Where the teams play each other and then the best team, in theory comes out of that but what they did differently this year, is instead of taking the top eight teams, they only take the top six teams and guarantee them a playoff spot. Then the bottom two slots are basically single elimination tournament between the seventh seed, the eighth seed, the ninth seed and the 10 seed fight for those spots. It’s led to some incredible games because they’re single–
Jon: You play wildcard.
Sandile: Yes, exactly. Because they’re single elimination. It’s like this super intense game the whole time and there’s really just no other avenue for that in NBA basketball because it’s mostly they take all the variance out by doing these really long best-of-seven series. Sort of the best of both worlds. I’ve been having a lot of fun with that.
Jon: That’s awesome. I feel like every major sports league has had to get real creative with structure and rules and it’s leading some really, really cool outcomes. I honestly think it’ll be a net good for forever, right?
Sandile: Yes. The creativity brought about the challenges of playing sports in a pandemic, I think will be a net positive for me in the sport I don’t matter really anybody else but I’m having a great time.
Jon: That’s all that matters, man. Awesome. Well, thank you so much. I certainly learned a ton by talking to you and I look forward to seeing your mysterious product at some point in the future.
Sandile: Yes, hopefully, it actually ends up existing. That’d be great. I know, this has been a lot of fun, John, thanks for having me.
Jon: Awesome. My pleasure, man. Happy hacking.
Sandile: Yes. See you.