Engineers at Remix are working at the forefront of mapping algorithms to build more livable cities
The SF-based startup is making public transit faster, more efficient, and more accessible in cities around the globe
For this story, we talked with each of Remix’s six engineers about the problems they’re solving, and how they work to fulfill a civic mission within a venture capital-backed company. In this related post, we spoke with Remix’s founders about the company’s origins at Code for America and their vision for livable cities. Also, the engineering team is hiring :) Contact the team about that: firstname.lastname@example.org.
What was your background before Remix?
I’ve always focused on social impact products. Before Remix, I worked at political, education, and fact-checking startups. I want my work to have a positive impact. That’s why I ended up here.
Tell us more about what brought you to Remix.
Given my background, I had a long list of requests for potential employers; Remix checked a lot of boxes. Most importantly, its mission focuses on social good. I was also impressed by the quality of the engineering here. At interviews, I would ask to see companies’ code. I came in with certain expectations, but I was really impressed. I also saw how well their products were selling, especially considering how hard it is to sell to government.
I liked the design too. It’s clear that two of the co-founders are designers — it’s beautiful and easy to use. Transit has such complexity to it, but the interface distills that complexity into something simple, which is hard to do.
Something I didn’t realize when I was interviewing, but appreciate now, is our close relationship with our customers. If we’re introducing a new feature, or we implement a brief shutdown, we just call them. They’re always happy to hear from us. We’re not the usual corporate types in suits trying to sell something. We are transit planners selling to transit planners — our products are tied directly to our mission.
“It’s clear that two of the co-founders are designers — it’s beautiful and easy to use.”
And, of course, I just enjoyed the people here. I had a number of interviews at other companies after I visited here, and I kept thinking about the Remix people. I knew even if I didn’t work for them, I’d want to spend time with them.
How will the future be different if Remix is successful?
Even with self-driving cars on the horizon, and even if we imagine smaller, faster, computerized, super-efficient cars, the problem of traffic isn’t going away. It’s not geometrically possible to move all the people in one city in individual cars. Efficient transit networks are essential.
There’s also the idea of open data. While some of the data we use is already publicly accessible — bus schedules, census information — we’re also gathering data about where people live and work and the percentage of low-income people without cars. We’ve included all these data points on our maps. It’s exciting to think about opening that up for other people’s use.
Improving transit systems contributes to energy efficiency and, we believe, financial equity. A lot of people can’t afford their own vehicle. Public transit gives them access to jobs they couldn’t otherwise reach.
What kind of challenges are you tackling day to day?
One thing I didn’t expect when I joined Remix was how many hardcore computer science problems we run into. We’re dealing with classic optimization problems and constraint satisfaction problems, like: How do you use as few buses as possible to cover the distance you want, or the number of trips in your system? Part of our product is figuring out how you provide the most service possible with the resources available. It’s really interesting to take classic computer science algorithms and use them in new contexts.
“One thing I didn’t expect when I joined Remix was how many hardcore computer science problems we run into.”
Recently, we had to solve the “shortest common supersequence problem” for a particular case; we were looking for existing solutions to that problem. There really aren’t any because it’s an NP-complete problem, which means you can only get approximations. But all the approximations we found were from bioinformatics, which is a very different data set than we have. So we had to come up with our own, brand-new solution that ended up working really well for our type of data.
Another challenge is presenting detailed geographic information with clear visuals. One thing we’re working to improve is how to visualize, for instance, 10 bus lines operating on the same street.
Has Remix changed the way you think in any way?
I knew nothing about transit when I started. Now it’s always on my mind. Previously, I didn’t think twice when taking the bus; you take it for granted. Now I know exactly how much work goes into it.
Tell us about what you were doing before Remix.
I just started six days ago. Before that, I worked as a data scientist for Clover Health, which is a health insurance startup. I think the common thread between Clover and Remix is that they’re mission-driven companies dealing with infrastructure that’s sort of at the backbone of society. At Remix, I’m in more of an engineering role than a data science role — I work on the algorithms behind our platform. I think what’s appealing to me in both cases is engineering a system that sits on top of some sort of algorithmic intelligence, whether that’s identifying patients that need a reminder to take their medication or identifying the most efficient way to staff a set of transit routes.
What compelled you to join Remix?
A combination of things. I was really interested in transit; it’s always excited me. I also get excited about the particular strain of algorithmic problems we work on here. And I wanted to work for a company that had a shot at making things better for those outside the affluent, tech-savvy cohort that a lot of tech companies cater to.
The fact that Remix even existed was surprising to me. It was exactly what I was looking to do without knowing the opportunity existed. It encouraged me that in 2017, tech startups can still contribute positively to transit.
“I wanted to work for a company that had a shot at making things better.”
Does the future look different with Remix in it?
Definitely. A better future with Remix means better transit for people who need it. That means the best and most frequent routes. Remix focuses on both helping transit planners discover new opportunities, and reducing costs so they spend resources more effectively.
What’s different about working at Remix compared to other experiences you’ve had?
One big thing is how thoughtful the management team is. They know each employee well, and they’re willing to put time into figuring out the best way to organize the company — to foster a culture that values both productivity and employees’ happiness.
What were you doing before you joined Remix?
Before Remix, I was at Google, where I worked on the Google Maps app for about five years. I also worked on the products Google offers for small businesses, as well as Google Search.
I came here because I wanted to work on something more civically focused. Google builds great products, but I wanted my job to directly affect a problem in our country and around the world. I’d read about upward mobility, and one of the things I learned was that access to transit has a significant impact on a person’s ability to improve their circumstances over time.
For example, if you live in a neighborhood where you don’t have access to transportation, it’s more difficult to find a job or go to school. As I learned more about the complexity of transit systems, I was shocked there weren’t better tools available for transit planners.
That’s where Remix comes in. We have great mapping technology that directly touches millions of people’s lives. We create tools for one transit planner who can then rethink and transform how transit operates in their city.
We hear you also built a Kanye emoji app?
Ha. Yeah. It started as a fun side project with my friend Chris. We thought, “Oh, wouldn’t it be funny to have Kanye emojis?” Apple had just updated their iOS SDK with some cool new features, including the ability to make custom keyboards. We built a simple app that lets you send Kanye stickers in conversations. Once it was in the app store, it got a lot of buzz. Then, Kanye’s people contacted us — apparently, Kanye didn’t think it was funny. We told them we were just fans, no hard feelings. Then the Kimoji app launched, like, months later. We’re not saying she copied us, but …
What are you working on at Remix?
In my role as a product designer, I oversee the functional and visual executions of all our products. Currently, I’m the only product designer at Remix, which means I collaborate across the company; for example, with our growth and success teams. I also work with our engineering team to design and build our products. We research the needs of our customers, which, combined with market research, helps jump-start our designs. Then we prototype and iterate various solutions.
“Access to transit has a significant impact on a person’s ability to improve their circumstances over time.”
At Google, I was working on a consumer product, which makes it easier to understand your users because you yourself are a user. Designing for enterprise products is different since our users have very specialized roles that require years of training. But both situations require you to put yourself in your user’s shoes. To do that, I’ve had to learn a lot about urban planning and transit design, which has been fun. It’s broadened my understanding of how cities are made.
What’s different about Remix compared with other places you’ve worked?
People say we have a low-ego culture, and I’ve found that to be true. People here are motivated by our mission. We’re all working on the same problem, on the same team, toward a common vision. Day to day, we’re focused on technically complex problems, but we’re always aware of the impact we’re making.
What are you excited about as Remix looks forward?
I’d like to push the envelope to make the urban planning process faster from idea to execution. Right now, it’s hard for a city to adapt to big changes like new economic influences or immigration. The faster and more efficiently we can update our solutions, the easier it will be for cities to react as their communities change and grow.
Tell us about why you decided to join Remix.
I worked for a railways consultancy in Melbourne, which meant I had experience in transportation. The job posting offered work in that field, but also had some challenging math problems. Solving them required drawing from my undergraduate studies. That excited me.
And now that I’m here, I’ve realized the engineering culture differs from my previous jobs — it’s more precise and more sound. The team cares about process; we test and review our code. We are proud to work with the code base we’ve created, and that helps us do our jobs without getting lost in the weeds.
What are you working on right now?
Because we’re still growing, everyone wears various hats here, but we also focus where our experience lies. Right now, I’m helping improve the speed and accuracy of one of our newest tools. That might sound dry, but there’s interesting, hairy math behind it.
At its bare bones, it’s an NP-complete optimization problem — meaning the universe has told us it can’t be solved. Nobody has found a solution that’s fast or consistent enough. So we trick the universe into letting us find an answer. We’ve thrown all the tools in our toolbox at the problem. Every day it works a little better. That’s part of why I’m here. I’m always looking for the interesting mathematical problem. I want to find and solve it before anyone else does.
Is there something at Remix you’re particularly excited about, or looking forward to working on?
Every once in awhile I have time to take on a compelling side project. Recently I’ve been working with demographic statistics. Right now we use off-the-shelf tools to calculate the population within a quarter-mile of the bus stops along a given line. Those tools are solid, battle-tested, but we’re running into their limitations. The next step is to see if we can give those statistics a wider context — maybe within a quarter mile of all the bus stops in an entire city, or how many jobs you can reach within 60 minutes’ travel time.
What will Remix’s success mean for the future?
I’m focused on solving problems, but of course it’s rewarding to see the solutions you’ve worked on affecting real life. We recently visited our customer cities and watched buses rolling out that are using Remix. Seeing the rubber literally hitting the road is satisfying.
What were you doing before joining Remix?
I did a bunch of things before joining Remix. I worked at Amazon for four years on a few different teams. I left to become the lead developer for an online custom wedding dress platform, which I did for about a year. After that, I launched a blackjack app, Blackjack Genius, that taught people how to play blackjack and get the best odds. Although it didn’t make a lot of money, it did have a widespread audience — over 10,000 active users at one point.
Why did you decide to join Remix?
Remix offered a nice balance of interesting user interfaces and social impact that other companies didn’t provide. Even though we don’t have millions of customers, our subset of customers affects millions of people. We are one connection away from those millions.
Right now we’re building some complicated user interfaces involving interactive timetables and bus-timeline diagrams. We’ve overcome some interesting challenges with rendering tens of thousands of rows of bus timetables without slowing down the page.
What other interesting challenges have you solved here?
My first big project was automatically matching bus stops to new lines as users draw them. Remix is sketch-planning software: if you want a bus to go down a different road, you probably want to go a certain direction, using the bus stops that are established on one side of that road. But we had no good way of automatically picking the stops once you moved the route. So the big algorithmic challenge was intelligently picking bus stops from a database of tens of thousands of possibilities.
“Remix offered a nice balance of interesting user interfaces and social impact.”
We were able to come up with an algorithm that picked the appropriate bus stops about 95% of the time. First, we find stops near the line. From there, we narrow down the list by removing stops on the other side of the street. To do that, we maintain a list of which side of the road the bus drives on in every country. Buses coming from all directions tend produce clusters of stops, so the final step is running a clustering algorithm which groups nearby stops together and then picks a single stop from each group.
What have you learned so far from working at Remix?
I’ve learned that planning transit is really hard — much harder than I imagined. I had no idea that transit planners were turning to paper maps and manually driving routes in order to measure the travel times. By giving them better tools, we can dramatically speed up the planning process, which improves transit for everyone.
What were you doing before you joined Remix?
I worked on consumer products at Yelp for several years, and I needed a change. I wanted to contribute to tools that created more properly functioning cities. For a long time, this kind of technology was shunned because of its complexity. Now I have the opportunity to improve it and hopefully make people’s lives better.
As one of the newer engineers on the team, I’m everywhere. I go all the way from the frontend to the backend on a daily basis. Every day is challenging in that so much is new. I am learning constantly.
Is there something interesting you worked on before Remix that informs your work here?
People expect that everyone here loves public transit and, by extension, hates cars. Yes, traffic congestion is terrible and public transit is important. But I actually build and race cars.
Since 2010 I’ve participated in an endurance race called the 24 Hours of LeMons. That taught me a lot about endurance, which is something I bring to my work here. You work as a team to find a vehicle that was destined for the scrap pile, fix it up for $500 total and race it for 16 hours. If your car keeps running that long, you win.
That experience, among others in my life, taught me to keep at it. Planning and predicting potential failure points allows you to adjust, which leads to more stability and predictability.
And, just like with technology, no matter how much you prepare, things still go wrong. You have to expect the unexpected and be able to come up with rigorous, simple solutions on the fly. We do that on the racetrack and we do that here.
That’s really cool. Have you ever won?
The Bay Area is actually really competitive so it’s hard to win, but we placed fourth and fifth a couple times. I’m one of the Guinness World Record holders for having the most cars on the race track at the same time — 240 cars.
How do you think the future will be different if Remix succeeds?
My girlfriend recently tore her ACL, so instead of riding her bike to work she now takes the bus. Her commute takes four times longer than it used to. She’s become keenly aware of how complex planning a city transit system is, and how easily a small mistake can translate to an inconvenient experience for riders. SFMTA recently started working Remix, so hopefully that will all improve.
Has working at Remix changed the way you think about things?
It’s been inspiring to work with people who care about things that matter to the larger public. You have to care to be miles better than anyone else, and we have a group of people here who do.