John Fawcett, 35, started Quantopian, a development and trading environment for quants, two years’ ago on a laptop set up in his garden shed. Like any good entrepreneur, he pursued an idea that was eating away at him in his spare time and only quit his job once he realized the demand for the product. However, he doesn’t recommend going it alone unless you have the right support network.
Quantopian is building the world’s first browser-based algorithmic trading platform, giving quants the tools and support they need to learn, create, explore and reap the benefits of quantitative finance.
I started working as a software engineer before becoming a fundamental analyst, covering the enterprise software market at a hedge fund. I was miserable as an analyst, and the portfolio manager of our fund took mercy and asked me to co-found a software company called Tamale, which was content management for analysts.
In 2008, we sold the company to Advent Software. At Tamale, I was able to meet many different kinds of analysts, including quants. I became fascinated by their creativity, their work and their super-high performance systems. I decided that I wanted to build the online platform for quants and began working on Quantopian with Jean Bredeche, another Tamale alum, who’s now our CTO.
It wasn’t as much a choice as a need. I had this idea to crowd-source quantitative finance, build an open-source platform and let anyone try quant analysis. Once I thought of it, I couldn’t stop wondering if it could exist and if the idea would work. After hours spent at home and on the weekends, I did everything I could on my own. I searched for quants in my network, emailed them and asked them what they wanted from a platform. I reviewed all the existing systems I could find and asked quants about the systems they had built.
At a certain point, I had a vision for what the quants I’d met wanted, and I almost couldn’t sleep at night because I wanted to see if I had it right. One Saturday, I couldn’t stop thinking about it, so I cleared all the lawn stuff out of my shed, put my laptop on the workbench there and built a prototype. It was ugly and did almost nothing, but some of the key ideas were there. You could type Python into the browser, and it would run fake market data through and generate theoretical results. I did a half-dozen demos that week and got so excited by the feedback that I gave notice and started working full time on Quantopian.
If it is your first company, don’t. Tamale was a startup, but I was surrounded by talented and very experienced business people. I had outsized responsibilities for a 20-something, but I had partners who were experts in finance and business. One of the best parts of my Tamale experience was our first office space in a machine shop of an engineering and design firm called Boston Engineering. One of the Boston Engineering founders, Bob Trieber, took me under his wing and really taught me how to design systems, debug problems, build teams and ship product. Then another individual, Mark Rice, joined as CEO, and he taught me how to scale an organization, operate globally, develop talent, delight customers and make sales. I can’t even quantify how valuable Bob and Mark’s mentorship was for me.
With Quantopian, I knew I wanted to move from being CTO to being CEO. To bootstrap with my own capital though, I had to do some of the initial development work alone. I used that process as a way to draw interest from the best CTO candidates I knew, and my co-founder and CTO, Jean Bredeche, jumped in and helped me from the very first day in the shed. He told me what technologies to build on, convinced me to learn Ruby on Rails and even try MongoDB.
All of those were incredibly valuable pieces of advice, and they also helped Jean and I build a relationship as CEO/CTO before we were really doing it. After Jean joined, he aggressively recruited the best people from his network. We are about half-and-half Tamale and Hubspot alumni, and everyone has worked with Jean in the past. The common connections allowed us to bond very quickly and get to the fun part of creating product and community.
We have six people, and we plan to do another phase of team growth starting this summer. We are at a good size now to be very nimble in the early product stage. We have a very seasoned team – the average tenure on our team is about 10 years of professional experience, so we have been able to move very quickly. By summer, we anticipate several major new services to be added to Quantopian, and it will be time to move into the next organizational phase.
The three most compelling things to do are: become a Quant in Residence, or QIR as we call it, participate in our forums and contribute to our open-source backtester Zipline. These activities show me that you care about our field, and it gives us a chance to really assess the fit. We’ve identified several prospects this way.
We need a ton of both, so it is hard to choose. Our community is growing very quickly, so the load and strain on our infrastructure is growing too. We are feeling the need for more infrastructure and technology operations personnel, but the community is also gaining in sophistication. We would love to have more quantitative finance people on our team who can help us with the next level of sophistication to our products – and keep up with our community.
You have to love the dramatic ups and downs. One minute you’re celebrating because you set a new traffic record, or shipped a new feature, and the next minute, you discover there is a bug in the new feature, and your application is slowing down from the load. With a startup, every small move is magnified, and to survive in the long run, you have to enjoy the excitement. I’d boil it down to tenacious, a bit thrill-seeking and fun.
Most of the technology built for quantitative finance is home grown inside trading firms. Software is critical, but those firms are in the trading business, not the software business. Monthly performance is the focus, not the software. The technologists I know in finance always lament the rush to ship features to make an impact this month or this quarter. Some say they feel their software is a disposable good, and I think it is true in some ways. At Quantopian, we aren’t just building a backtester or a trading platform. We want to build THE backtester and THE trading platform. Much of our IP is in an open-source library called Zipline, which means any of our core engineers can walk out the door and still use their creations. We are building a community for Quantopian and an ecosystem for Zipline. There are very few opportunities in finance to do creative work for a massive global audience such as ours.