Tech vs. People: The Hardest Problem with Will LaForest | Ep. 26

Tim Berglund talks to Will LaForest (Confluent) about his career in software and data streaming. Will’s first job: a high school internship at DARPA. His challenge: turning advanced technology into something people actually care about.

SEASON 2
Hosted by Tim Berglund, Adi Polak and Viktor Gamov
Produced and Edited by Noelle Gallagher, Peter Furia and Nurie Mohamed
Music by Coastal Kites 
Artwork by Phil Vo 

•  🎧 Subscribe to Confluent Developer wherever you listen to podcasts. 
• ▶️ Subscribe on YouTube, and hit the 🔔 to catch new episodes.
• 👍 If you enjoyed this, please leave us a rating. 
• 🎧 Confluent also has a podcast for tech leaders: "Life Is But A Stream" hosted by our friend, Joseph Morais.

Transcript

SPEAKER_00 0:00

The technical challenges I never found to be the hard ones. It's always the human ones that I thought were the hard ones. For me, anyways. You need to be able to communicate, you need to be able to talk in human terms. But if you don't have deep technical understanding, you don't have any credibility.

SPEAKER_01 0:13

Hello there, everyone. I'm Tim Berglund, and welcome to Confluent Developer, the podcast where we explore the fascinating journeys of software developers tackling really hard problems. In this episode, I'm interviewing Will LaForest. He's a field CTO for Confluent. Will and I have worked together for, well, a long time. And today I got to talk to him about a high school science fair project he did, of all things. And you know, it wasn't even a science project. He just wrote some code he thought was cool. I really get that. Then he learned some really important lessons about how to bridge the gap between the technologist's fascination with things and how they work, to the needs people have in understanding those things in terms that matter to them. Those are lessons we all need to learn. Let's get into it.

SPEAKER_00 1:15

That's what they say.

SPEAKER_01 1:16

That's what they say. What what would they say that a field, what would you say that a field CTO does? For anybody who isn't familiar with sales organizations, what do you do?

SPEAKER_00 1:25

Yeah, well, I mean, it's a fancy title for saying that I spend a lot of time talking to customers, helping them sort of solve their problems, with of course, in in this case, Confluence. Um gathering signal from customers, like, what are the problems they're running into? What could we be doing better at Confluence? So obviously that's bringing bringing it back to product is part of that. Um and then I think just expressing what this technology space can do in terms of that matter to them, like in their business. Right. I think that's actually interestingly one of the sometimes the hardest things. Because you know, when you're talking to executives, as much as I you and I both love technology, right? Um, they don't necessarily know or even care what Kafka is.

SPEAKER_01 2:23

Maybe, maybe tech isn't their background. Maybe it's been a little bit, you know, and uh they're not au coron on everything that's going on these days. Yeah, yeah.

SPEAKER_00 2:34

So a little bit of this, a little bit of that.

SPEAKER_01 2:36

Um and that's the that's the difference. To be a field CTO, you have to basically have had a background as an engineer of some kind or a technical person in whatever space you're in, but you need to be able to articulate those things in terms of of uh somebody higher up the food chain.

SPEAKER_00 2:53

Yeah, exactly. That's that's sort of the axis of skills required. I think you need to be able to communicate, you need to be able to talk in human terms, but if you don't have deep technical understanding, um, you don't have any credibility, right?

SPEAKER_01 3:06

So yeah, right, right. Because there's other folks in the room who who will call you out. For sure. What uh okay, that's what you do now. Let's go to the other end of your employment journey. What was your first job?

SPEAKER_00 3:18

Well, um kind of depends on how you define job.

SPEAKER_01 3:23

Uh like so uh spicy.

SPEAKER_00 3:27

Go on. No, no illicit uh non-medicinal drug vending here.

SPEAKER_02 3:32

Okay, okay.

SPEAKER_00 3:37

But you know, when I was 10 years old, I I did like stuff uh flyers and people's doors for pizza delivery. So I did do that and I did get paid money.

SPEAKER_01 3:46

Cash money.

SPEAKER_00 3:48

I guess, but okay, so aside from that, I think the first, I guess, job where I actually had to figure out I was like, oh my gosh, who's the IRS?

SPEAKER_01 3:57

Well, somebody who's reporting to the IRS, yeah. There's a WT.

SPEAKER_00 3:59

Why are they taking money? Um, I was kind of uh I was kind of very um blessed, if you will. I had a paying um internship um at DARPA for two summers in a row, and that was that was fun time. So that was really good. I guess you could say that was my first paying job.

SPEAKER_01 4:19

Yeah, that was college, um high school. High school, okay.

SPEAKER_00 4:23

That would have been 90 uh 1991, summer of 92 and summer of 93.

SPEAKER_01 4:29

Yeah, okay.

SPEAKER_00 4:30

Yeah, um that's kind of cool. What what did you do there? Um well, this is this, yeah, this was formative for me. I um this is when I learned to uh program assembler, so that was that was fun.

SPEAKER_01 4:47

Nice. What processor, if I may?

SPEAKER_00 4:49

Um, well, in this case, it was an embedded uh risk-based processor.

SPEAKER_01 4:54

Okay, okay. I was also writing firmware in the early to mid 90s. That's a good okay.

SPEAKER_00 5:00

Wow, I had no idea. Yeah, so that was fun. You know, up to that point, I'd been doing like I'd done a lot of languages, Pascal and C.

SPEAKER_01 5:07

As one did at the time, yeah, as one did at the time.

SPEAKER_00 5:09

So yeah, Pascal was always in the computer science. Like, how do you learn program?

SPEAKER_01 5:12

You're gonna learn Pascal Pascal, and that just doesn't happen anymore.

SPEAKER_00 5:16

So that was fun. I I I fell in love with next uh computers there, next step, if you remember. So I did a little bit of Objective C. So yeah, I did a little bit of this embedded systems, and then I did some Objective C programming to do some sort of administrative task, nothing mission critical, of course, but it was good. It was fun.

SPEAKER_01 5:37

Nice, cool. Yeah, so between field CTO now and uh either paper delivery, uh you know, pizza delivery flyer boy or intern writing Objective C, uh, what would you say maybe one of the top most interesting problems you've solved?

SPEAKER_00 5:56

So yeah. So interesting problems. Um I was thinking, do I do I go the wood direction where I talk about the hardest problems or an interesting problem? Because they're not your call. Your call.

SPEAKER_01 6:09

This is your this is your uh interview.

SPEAKER_00 6:12

I'm gonna go, I'm gonna go hardest problems simply because I suspect you talk to a lot of really smart people that talk about like really technically sophisticated and interesting problems they solved. And I, you know, I I did some of those um for sure. I did some fun stuff um when I was at Mark Logic and stuff. But like I think you probably get enough of that in the interviews. So I'm gonna go to the opposite spectrum because I think when I was sort of contemplating in preparation for this conversation, so this is a this is a cheat for everyone watching out there. Yes, we there you go. I do you do think ahead of these things.

SPEAKER_01 6:49

Yes, yes, I I did at let Will know what I was gonna ask him.

SPEAKER_00 6:52

Yeah, exactly. Um I thought about this and I was like, you know, the the the harder problem I that I thought that I think can tie to my current job really came down to, I think, um, okay, so I'll set the scene. So this is, and I'm gonna roll back the clock um because I've seen this movie a few times, they're kind of related. I'm gonna roll back the clock to 1991. So I did um this was I was in high school, I did a science fair, and my project was um at the time I was really fascinated by they called it AI um back then. But essentially it was how do I play AI? And I was running, you know, 486 on DOS, and so yep. Um so my first project was I made in in what I called an AI uh baseball pitcher. Actually, a friend and I did.

SPEAKER_01 7:45

Okay.

SPEAKER_00 7:46

Okay, so that was pretty successful. That was great. So now fast forward to the next year.

SPEAKER_01 7:51

So that was just software. There was no robotics or anything like that.

SPEAKER_00 7:54

No robotics, yeah. And so, well, yeah, exactly. So I we simulated basically had different profiles of batters, and the trick was could we use essentially this was a single-layer neural network um sort of AI? Could we use that to try and improve our chances against these these batters?

SPEAKER_01 8:13

Okay, okay. So it was kind of fun. Certainly in the in the hacker or undergraduate space, you know, that would be yeah, hip and cool and new.

SPEAKER_00 8:26

Well, I mean, and the only reason why I even knew it, because back then, at least I so when I was at DARPA, obviously I got introduced to the internet, but prior to that, there was like Bolton Board systems and 300 bod modems and stuff like you couldn't just find this stuff on the internet. I just happened into a book at a university bookstore that was on this, and that's why I started reading. I was like, you know what? I'm gonna try this. So, so anyways, that was cool. But the next year I decided to take it to the next level. I really, and so this is gonna sound ridiculous, but this is so my friend and I decided, okay, can we apply this to like general knowledge, i.e., if someone asks a question, could we recall something? And so that sounds like the interesting challenge I want to talk about, but that wasn't. Um, and in fact, we gave up on neural networks, it was just way too hard. Instead, we just did like essentially some heuristic-based natural language processing. It was really terrible. Like if you use the wrong grammar, it wouldn't work at all.

SPEAKER_01 9:27

Eliza, Eliza level. That that stuff did not work then at all.

SPEAKER_00 9:30

No, it was terrible. I mean, it would probably have been bad for anyone, but we were high schoolish programming in C. So and we did like a knowledge graph. So that was kind of fun. But here's the challenge. I'm getting to the punchline. The punchline was the night before the science fair, we're putting the finishing touches on it. I don't even remember what happens now, but the but the bottom line is we lost everything. And like we had three and a half inch floppies, and we like try, I knew it was code was on one of these things, and we just couldn't find it. And so we reached, we were in this point. Like, what do we do now? Because clearly what we did the previous science fair was to sort of show the learning in process to show how like the pictures were getting better against batters. And we could actually demonstrate this for the judges. Back then, your science judges were not computer people, like they they're not gonna.

SPEAKER_01 10:22

And if I may, so pet peeve raising my own kids, they were homeschooled, but there'd be like an enrichment program. There'd always been a science fair because on their dad, obviously that's gonna happen. Yeah, and science fairs are usually uh that there's this this big push to use the scientific method, which I happen to love the scientific method. It's an amazing way to discover, to create strong knowledge claims about the you know physical reality. It's great. Yeah, um, it's not always very fun, right? Like, do you want to do an experiment or do you want to do some cool thing? Like uh, you know, that that makes kids say, Wow, that's really neat. Uh what's that all about? Which I always just did like you know, really an engineering fair kind of project or something. Kind of sounds like what you're doing. You're like, science, whatever. Okay, I'll contrive an experiment. I want to put a neural network. So here we go.

SPEAKER_00 11:11

That's exactly right. Yeah. I I don't remember anything about the scientific method in that science field. If we're gonna be honest, I kind of ignored it. But you're right. It is the like finding your passion and your love, right? That's what it comes down to, I think. Um so I think that moment was like, what are we gonna do? Because we had worked for, I don't know, we actually started immediately after the last fair. We had worked like most of a year trying to put this together. And it and we had a demo that worked, and it was really cool. You could like it would ask some questions, you would answer some questions, and then it would allow um you to ask some questions based upon that knowledge. And of course, we had a happy path demonstration for the judges that we could do. But it all fell apart. And so the question was what, but like we couldn't just pull out predominantly awesome. One of the reasons, like I kind of wanted to give up, but like my um my aunt had made this super fancy science fair backdrop for me. So she it was like hand painted and it was incredible. It was like much better than I deserved. And I feel guilty about this ever since. But like, I was like, well, I can't pull out. She did all that work. Like, I gotta do something. So the challenge was okay, well, how can we demonstrate it to them? And or how can we put it in terms where they can understand what we're trying to do here with natural language? I don't even think I called it natural language then, to be honest. Um, but uh like this this graph. Um, so we tried to build a demo, we couldn't get it done like a fake one. We were gonna say, we did the real work, but here's a fake demo. Right. Ultimately, it failed. Like we just did not have something we could demonstrate at the science fair. And so when the judges came around, what I tried to do was explain all the work we had done for the that year in terms they would understand. Um, and the punchline was we failed at that too. Like they just didn't get it. And so, like that probably was the single most difficult moment of my technical career. But I would say what I learned from that, and and what we did was so much more ambitious than the year before, to be honest. And what I learned from that was like trying to like how valuable it is to be able to express some technical approach in terms that your audience really can relate to and then they can see. Because, you know, I felt like if I had if the same thing had happened, and because and I did not do that with that Second Science Fair. Like we, I just I should have done that. What I should have done was thought about what this meant in something that everyone can relate to instead of trying to go into, oh, well, we baked this build, you know, we had these nodes and we have these edges that link them and we could traverse the graph. Instead, I should have done something, some terms that they understood. So I've seen this play out a few times in my career, but that was the worst moment. It was like, oh, what are we gonna do? Um, but it was it was maybe formative in why I got to where I am, I guess.

SPEAKER_01 14:30

I I totally see that. Now a quick word from our sponsor. Confluent Developer the Podcast is brought to you by Confluent Developer the website, which has everything you need as a developer of data streaming systems. And it's completely free. We've got curriculum, hands-on exercises, executable tutorials, the online data streaming engineer certification, also free, a way to find a meetup near you, those are free. Everything is there. I really want you to be successful in your journey as a data streaming engineer, and this is the site that has what you need. Check it out at developer.confluent.io. That's developer.confluent.io. Now back to the show. Because so a lot of the times when I ask folks this question, I'm I'm like trying to dig into what was wrong with the world before you went down this path, you know. Well, in your case, you were curious. And and you know, the first year you're like, you know what would be awesome is if we wrote a neural network in C. Yeah. Um now, so most people, and I don't think any of them are listening to this podcast unless it's in the car and somebody are, Tim.

SPEAKER_00 15:41

Whoever, I don't know who people you're talking about, but I'm sure they're listening.

SPEAKER_01 15:44

Well, the most people don't think what I was gonna say is don't think that sounds awesome. Okay. That's not typical. That's the thing you and I have to recognize. That yeah that I mean, because I'm I'm with you. I'm like, that is pretty freaking awesome. Teenage will, like neural network, next year, knowledge graph, natural language processing. I'm on this because it's cool, right?

SPEAKER_00 16:06

Yeah, not back then it wasn't.

SPEAKER_01 16:09

Well, no, it it that was just before the internet, yeah. So you and I, um just a couple years older, and that was definitely uh no, it was not cool to be into computers in the 80s and early 90s. That uh, you know, uh 10 years later, a different game. But um you were curious, you were amazed by the fact that those are things you you happened upon a book uh uh book describing technologies that in a much more developed and uh you know compound form uh changing the world right now, um, you know, through through uh the transformer architecture and large language large language models, yeah. Changing it or turning it upside down or some some some some version of that. But you just thought that was awesome. There was no problem. There wasn't anything that pushed you on a journey away from home other than that's dope, I'll do that. And I just want to highlight that because that kind of curiosity uh isn't usually like when you're talking to a customer, I'm gonna guess that's not usually what starts a customer on some initiative, but it is what starts people discovering things. So that's a different kind of beginning. I I'll let you respond.

SPEAKER_00 17:27

Yeah, it's true. I mean, in hindsight, like that sort of approach is like kind of what I would never recommend a customer do. Uh you know, right? Because I was like, oh, I had a something I want to do looking for a problem rather than the inverse is like, oh, I know I have a business problem or you know a challenge, I need to solve it.

unknown 17:50

Right.

SPEAKER_00 17:50

But I guess that is kind of the point of like academia to some degree, um, or RDs, like you know, you don't necessarily have to start with a problem. You can start with the creative exploration side, I guess.

SPEAKER_01 18:04

So you can start with you know what would be cool. And the the contrast is important because you're right. When when you're talking to a customer, hopefully that customer is not starting with, you know what would be cool, right? They they need an outcome, they need to make money, save money, mitigate risk, they need to know what the numbers are, and and you know what these objectively observable things that they're trying to accomplish are in a business, because they're responsible, they're agents, and they're responsible for for managing the resources of the principal, of the owner, yeah, of whoever the shareholders, whoever that is.

SPEAKER_00 18:42

Yeah.

SPEAKER_01 18:43

All kinds of they they that's what they do. And so you now talking to customers, there's this like vicious utilitarianism that that you you have to truly for valid reasons, you you have to respect. Um, but I mean, behind me, if you're looking at the video version, wrong way, there's some shelves back here, and you can't see there's some electronic test equipment and stuff. I like to make stuff because it's fun. Like, did can are there easier ways to solve all those? Yes, but it's just kind of cool, you know. And and that's a separate thing from the the quasi-selling that you do. But I just I love that about young Will. That's that's all I want to say. He was good. He wanted to do something. Yeah, it was and you already said this, but but he learned a lesson. Uh talk about how maybe give us an example of how you see that play out these days when you should have been able to describe this deeply technical thing that nerd you cared about, and these teachers had no idea what you're talking about. You should have made it been able to make it relatable. Um yeah.

SPEAKER_00 19:52

Yeah, I think um for me, the way it plays out now, um, and and by the way, I don't always do this. Like, you know, sometime I I I get wrapped up in myself or some technology or or whatever. But, you know, in general, what what it's taught me is like the first thing that matters is the first thing to realize is like the person you're talking to is who matters, right? So understanding them, understanding where they're coming from, and then relating ideas um with a narrative that they care about, like story, tell a story, storytelling. Because I'm sure you know people have said this. I'm sure you've said this. Like we've been telling stories for, I don't know, since we've had 100,000 years or something like that.

SPEAKER_01 20:42

Yeah, since we've had words.

SPEAKER_00 20:44

Since we've had words. Um and and I think so, so a couple things I trying to do is like I it's it is intensive because you know, I I talked to literally hundreds of customers in a year or certain, certainly triple digits. Um, and so you don't, you know, prepping for each one and trying to understand where they're coming from, what they're you know, the what their business does. Like, even though I work across a lot of industries, I'm telling you, it's hard. It's like there's only so much it can fit in your brain, right?

SPEAKER_01 21:14

So I don't, I mean, you you do customer meetings for your job. Yeah, um in my you know work responsibilities. I I do them sometimes. And I love them because it's this constant struggle to understand. Like maybe you have an hour and you are just kind of flying through the conversation. Number one, trying to understand their business. Can I wrap my mind around what it is you do in enough detail to think about the technology stack and and whether the stuff I've got in my bag here is any good to you? You know, I so I love that problem. Yeah, you're trying to understand the business. Sorry, go on. You're trying to explain it in ways that they can understand.

SPEAKER_00 21:58

So, I mean, that sounds Like 101, but I think it's something to always be mindful. And I guess this the second thing that I try and do whenever possible is I I try to have uh a demonstration on hand that I can use to actually show some of these things that's almost demo domain oriented. So it's not always possible. You can't always have a demonstration of said, you know, some technology, in this case, confluent, ready that shows, you know, how I am, you know, optimizing supply chain for retailers.

SPEAKER_01 22:33

Like yes, or manufacturing process or pharmaceutical or or a little of a you know, whatever. You do it's not always domain specific, but it needs to be close for this executive person.

SPEAKER_00 22:43

Exactly. That's right. Because you otherwise they'll fixate on the wrong things. So I that's the other thing is like I I try because I remember like doing like explaining my science experience, you know, experiments using like uh Aldislamp or uh what's the what's the flags they use? Uh yeah, um that's not fun. It's much better to show it. It's much better to show that that AI pitcher pitching baseballs. Um yes. So yeah, I get I think those are the things that I I kind of like. I failed miserably. Um, and I learned demon, you know, allowing people to see things in action is really, really powerful, specifically if it relates to them and understanding their perspective and where they're coming from. You know, technical problems are I love it. Like they're you know, when so when when I knew what I was preparing, I was like, oh, there were so many projects I worked on that I really, really love that I could talk about and geek out on. But like honestly, you know, if you have enough time and you know, you're willing to put some sweat into it, like you can do some amazing things, I think. Um, so like the technical challenges I never found to be the hard ones. It's always the the human ones that I thought were the hard ones for me anyway.

SPEAKER_01 23:52

So and I think this is a key. This is just kind of coming to me now. So I I you're helping me realize some things. I love this conversation. Um the reason why we have technological advancements of any kind. I mean, you and I care mostly about software. There's all kinds of other technology. You know, there's metallurgy, there's there's chemistry, whatever. Pick the thing you like. Um the reason why we have those things is because somebody at some point said, huh, you know what would be cool, or oh, that's weird. And and they wanted to dig into it. They wanted to do something, they wanted to understand it. And they were very focused on the thing. And that's not that's not a people problem. They're focused on the thing. And that's you picking your supposed science fair project, really an engineering fair, because that was more awesome. It's what you wanted to do. You're not thinking about people, you're thinking about the thing. And that uh that sounds like a negative framing. It's great because you have to think about the thing. There has to be somebody who wants to do that, otherwise, there's no technological innovation. When it comes time to explain that thing to somebody, like if you happen to be selling it and you're interested in helping somebody make the decision to buy it, well, now that's a people thing. And you have to explain it in terms of them. You have to understand them and tell a story that's relevant to them, which is not the same as let me tell you how awesome this thing is. You know, yeah. For you and me and anybody else in this business, you can stop there. It's awesome. You just want to think about the thing. Yeah, you have to be, and this is something I think magical about the the pre-sales slash field CTO kind of world. Um, to some extent, DevRel, what I do is is a little like this, but I think it's a bigger deal for you. Uh, you you gotta you gotta be able to know the thing, which means you're intrinsically interested in just geeking out on the thing, but then you you gotta be able to to pivot to people. You you gotta be able to do both at the same time. It's kind of a little magic there.

SPEAKER_00 26:00

Yeah. Um it is. I love it. I'm not gonna lie. Like people ask me, you know, what is it that I feel CTO does? And then I describe it, and like I'm not gonna lie. I I have a great gig. Like, it is a lot of fun. I mean, I imagine the same thing goes for you, as much as I love my job. Like, there's some things, you know, that I you know, maybe I'm doing XYZ, I'd rather be doing something else. But on all on whole, it's an amazing job. Like, I feel very blessed.

SPEAKER_01 26:29

Uh as they say, the lion does not concern himself with with Slack notifications and and you know, but uh yes, it's that it's amazing work that we get to do.

SPEAKER_00 26:39

Yeah, and and you know, it's the other thing I think it's great about it, and I always say this to people like when you're thinking about where in an organization like this you're working, the other really great thing, it it doesn't have to be field CTO, you know, but like when you're taught working across a lot of customers, you just learn a lot of interesting things about businesses and meet a lot of really cool people. I mean, that's a lot of fun too. Exactly, yes. Whereas like, so so that's for me that was always the tension. I love building things and I love seeing it come to fruition. I just love that. Like, oh man, I feel so good when I open up uh an IDE and start coding. But when you do that, when you're focused on from cradle to um, I'm not gonna say grave, but cradle to delivery.

SPEAKER_01 27:29

Yes.

SPEAKER_00 27:30

Um, it takes a long time. And um, and so it's hard to get sort of the breath. So it's like a sacrifice, I think, for me, anyways, because I love both aspects. I'm sort of sacrificing the engineering side, which I love so much. Um, but it means I get to talk to a lot more people and hear a lot of interesting problems and and stuff like that. So that's sort of sort of the balance. And I I think people just need to think about that. Like, what do they love more? Right. Um, because it's important to love your job. But anyways, I'm rambling.

SPEAKER_01 28:00

Um, good distinction and I think good challenge. What what thing do you love more? Uh, what balance do you want between thing and people? And that helps you position yourself in the whole spectrum of tech careers.

SPEAKER_00 28:13

Yeah, definitely. Yep.

SPEAKER_01 28:15

My guest today has been Will LaForest. Will, thanks for being a part of the Confluent Developer Podcast.

SPEAKER_00 28:20

My pleasure. Thanks for having me, Tim.