By Brent Goldman - March 18, 2021
I’m excited to introduce our newest member of the Sisu Engineering team, Nathaniel Tucker, as a Staff Frontend Engineer on the Product Engineering team!
Nathaniel joins us most recently from Coinbase, where he was one of the first 100 engineers and focused on improving both the product and developer velocity. Prior to Coinbase, he was the Lead Web Engineer at Robinhood, where he grew the team from zero and built the trading and stock exploration web platform. Nathaniel is particularly skilled in architectural efforts to increase product quality and developer velocity for rapidly growing engineering teams. We’re excited to have him join the Sisu team to help accelerate velocity and quality here.
To properly introduce Nathaniel, we sat down for a brief interview to learn more about the challenges he likes to tackle, why he’s excited to join the team, and his thoughts on philosophy, self-driving cars, and the great state of Texas.
I was interested in Sisu for two reasons. First, Sisu is at the forefront of exciting challenges. With Snowflake making data collection so ubiquitous, Sisu is tackling data empowerment’s next frontier — making complex data and its implications clear. I wanted to be a part of solving that. Second, I’ve known you for a long time, am impressed by your empathic insight and thoughtfulness around engineering culture, and was excited at the opportunity to work with you and the team you’re building.
First, I think there’s an interesting challenge in increasing the developer velocity at Sisu while at the same time building a high-quality product with a really lean team.
At Coinbase, one of the more interesting problems I worked on was scaling the codebase. In doing so, we had to deal with many different complexities and determine how we could apply those same principles around the company so that other teams could take advantage of the architectural wins we developed.
Compared to Robinhood and Coinbase, what’s unique to Sisu is that the problem we’re trying to solve requires processing and interpreting a lot of data. This creates challenges around taking enormous complexity and presenting to users in a clear and actionable way. I think that this will require a lot of iteration on the product design, as well as the client infrastructure supporting that. One of the powerful ways to take on this challenge is by increasing developer velocity to iterate and get to new and innovative solutions a lot faster.
Inevitably a mix is required, with weight based on how well understood the patterns are. Doing feature work allows discovery of these patterns naturally, which can easily be iterated towards reusable patterns. Starting with frameworks can be more productive if the patterns are well known, as their benefits will be exploitable more quickly. However, this pattern can be very dangerous for less known patterns, as building dependencies on these APIs can create friction around any changes. In either case, it is imperative to keep initial designs as simple and low level as possible; optionally following up with specialization layers.
Part of the re-architecture at Coinbase was what I call the modern Flux architecture, and one of the core pieces of technology is this architecture utilizes Rest Hooks.
When I started two years ago, React hooks were still in beta release, and this meant not only did I have to advocate for a completely new architecture, but I had to do it with a technology no one fully understood. It ended up being extremely powerful in improving both our developer velocity and the code quality. One of the problems we were experiencing at the time was playing a lot of bug “whack-a-mole” — we’d fix one thing and then find it causes another problem. Addressing this meant we had to restore the data integrity first, and then we could improve performance.
Rest Hooks was a foundational effort to address these problems because much of our application complexity was due to networking data. Rest Hooks solves issues of consistency, type safety, and performance in a systematic way that enables developers to iterate a lot faster, automatically solve a good chunk of these problems, and ultimately allow our site performance to go through the roof.
When I first introduced these at a hackathon, our team built what we called “Coinbase Lite,” which was basically a re-implementation of the entire site in one and half days – all using React hooks . It was an exciting demonstration of what we could do with React hooks and how fast it enables developers to produce things. At the end of the hackathon, when we displayed the side-by-side of the current website and the one we built with React hooks to the CEO, he was speechless and only said, “Wow.”
In general, I enjoy understanding things by taking a “scientific method” approach to life and using that understanding to predict things and what the implications of these things will be. Some people refer to this as intellectualism, but for me, it’s a mode of operation to explore ideas more deeply.
One of the things I’ve been thinking about lately is how writing can convey complex and challenging topics in a way that’s accessible for multiple audiences. One of the advantages of conversations is that it’s typically between one or two people, and you start from a basis of understanding that you can build off of. But when you’re writing, you have a much broader audience with people of different backgrounds and contexts, so determining how to relate things in writing becomes a challenge.
What I’m struggling with now, in both technical writing and philosophical writing, is how to explain a concept with one level of abstraction and using the web to link people who might not have a background in the topic to the knowledge and resources they need to learn more.
It’s a little less controversial than when I first professed it in 2017, but the level to which I am bullish on Tesla is probably still somewhat controversial. I was actually company-famous at Coinbase for being a big Tesla advocate. It’s not just succeeding as a car company, but I think they’re going to take over a large portion of transportation and energy – which if you look at the economy of the US and Europe, one-third of the economy is energy and another is transportation. So, if Tesla can get even 5% of both, that’s an enormous piece of the pie.
The controversial piece is their approach to autonomous driving. Where all other companies use LIDAR, Tesla is famous for not using LIDAR and what’s interesting is that their strategy is actually making money with what they’re producing today. Meanwhile, other companies are burning money and will really only create a taxi service — which is not where the money is in transportation, it’s in trucking. Entering that space with a Tesla Semi truck coming out will leapfrog them into sort of dominating the autonomous space.
Ah, well, there’s where you want to work and where you want to invest your money. I’ve been traditionally bullish on Amazon as well, but I would never work at Amazon – and I’d never work at Tesla. There’s a lot of factors but, most of all I value my work being a core business advantage, as well as having “fun” at work. I don’t see my skills being key to the success of either Tesla (electric drive-train, autonomous driving, manufacturing) or Amazon (supply-chain efficiency, server infrastructure).
I was born and raised in Texas – and currently live in Austin! At Robinhood, I randomly brought up two different Texas facts at a meeting. After that, a lawyer there asked me to share a new Texas fact every day. I took the challenge and started looking up all this stuff and learning more about Texas.
One of my favorite facts is that the original Six Flags is from Texas, and their name is shortened from “Six Flags over Texas.” It refers to the fact that Texas had six different flags, or six different nations, controlling it throughout its history.
Want to build product features with high velocity based on Nathaniel’s frameworks or over-analyze Tesla vs. Amazon as the world’s first megacorp? We’re hiring Frontend and Full-Stack software engineers, as well as across the stack in engineering. Come join us!