Olivia (right), Josh (left), and I (middle) at a Santa Monica Pier Concert sponsored by Snap, Inc.
Introduction
Two months ago I graduated from UCLA with a bachelors of science in Mathematics/Economics. During my junior year I founded Flux Ventures, a company that created Flux Chargers, an e-commerce portable charger brand, and Flux.LA, a digital marketing and SEO firm. Due to the revenue from these businesses, I did not have to work a traditional internship towards the latter half of my college career.
As my senior year came to a close, many students in my classes sought quantitative analyst positions at large firms. However, since I had the luxury of working for myself while at UCLA, I wanted to pursue some new projects with the Flux team. We had largely automated our chargers and digital marketing businesses, which provided us with a decently sized runway to explore new ideas. This was extremely liberating, as we were not inhibited by the necessity of making money.
Recruiting a Team
Through Tech.LA, I interviewed 4 highly qualified computer science students from some of the top universities in the world to work for me full time for summer. We were looking for an iOS engineer to help develop an iPhone game and a machine learning engineer to help craft a neural net to predict fluctuations in cryptocurrencies.
On behalf of Flux, I extended 2 offers. They both accepted.
I hired Olivia Brown, a Computer Science student at Stanford, and Josh Miller, a software Engineering student at RIT. I secured a short term lease in a two bedroom/two bathroom apartment near UCLA. Olivia and Josh lived in the respective rooms and we turned the living room into an office, with desks and monitors bordering the entire room. My co-founders Alejandro Rioja and Scott De Taboada outlined and managed these new projects with me.
iOS Application
Prior to working for Flux, Olivia created Lippi during a hack-a-thon. The iOS app disrupted the beauty tech industry as it served as a hub for lipstick discovery and sales. She negotiated a commission for every referral Lippi made to a company’s site, and amounted 300,000 downloads. Better yet, she made this at the age of 16. I had never seen anything quite like this on a resume.
I appointed Olivia lead developer of Poos Caboose, a viral jumping game for iOS. The idea was simple enough that I felt we could push a product to market within the first month of summer. With myself as lead designer, we started in June and released on the App Store in July. Due to our use of Apple's haptic feedback in gameplay, we were featured on the App Store in 155 Countries and were the top up voted product on Product Hunt upon release.
Adversarial Neural Network
Prior to working for Flux, Josh was extremely active in the machine learning community as a self-taught machine learning aficionado since the age of 19. His GitHub even features some of the first applications of concepts discussed in machine learning papers published days before his implementation. Yet, Josh’s resume caught my eye due to a project he listed called CooperBot. It’s a Twitter bot that uses a neural net to produce tweets like Anderson Cooper. Naturally, I thought this was hilarious, so I knew we would get along.
I had Josh start out by scraping Bitcoin, Ethereum and Litecoin price/volume data to backtest trading strategies. The intent was to eventually develop a neural net to predict price changes and create a trading algorithm that could profit from these predictions. We started by using market orders, but opted to limit orders to avoid the 0.3% commission - especially due to the frequency at which we were trading.
By the third week of summer, we were managing a sizable fund of own money in real time using the GDAX API, and by late July, we were profitable. To shed light on the frequency of our trading, the total volume of our trades exceeded $200,000 despite our fund only being a few thousand dollars. We were literally trading almost every second. We named our bots, Bernie (because he trades Bitcoin), Eddie (because he trades Ethereum), and Larry (because he trades Litecoin).
The Future
Even though Olivia and Josh return to school in September, Flux will continue to develop both projects on their path to monetization. For Poos Caboose, we plan to add complexity to the gameplay, a social leaderboard using Facebook’s API, and a handful of new hilarious Pooses to optimize user retention. With the adversarial neural net and proprietary algorithm that Josh made, we plan on raising money and starting a cryptocurrency hedge fund.
Looking back, I am very thankful Flux pursed these new ventures, as it was the only way that I, a 22-year old recent graduate from UCLA, could gain experience managing and leading two highly technical projects in fields that directly interest me. Instead of looking for jobs, we created them.