An interview with Kostas Manolarakis, lead research and development engineer at stratagem technologies
Read more on what Kostas Manolarakis has to say.
1. Tell us about your career. Where did your interest start? How is your career unfolding?
Ever since high school, I had a keen interest in mathematics, and was always drawn to working in a quantitative discipline. I’m very lucky to have grown up in an era where quantitative industries are growing fast and these types of skills and placements are much sought after, compared with twenty or thirty years ago where options were limited.
My love of mathematics in high school led me to study this further at Cambridge University and undertake a PhD in Numerical Stochastic Analysis at Imperial College London. It was there I realised that the pure research roles found in academia weren’t the right fit for me and so I made the transition to the finance world.
Having studied stochastic analysis, the world of derivatives felt familiar, however I was always very interested in data driven roles. When the opportunity to explore this at Stratagem came up, it was a great match and excellent place for me to apply my skills to a new problem.
2. Who are Stratagem’s customers?
We have a number of different customer groups ranging from B2B to B2C. We have been working in R&D for several years to build an analytical backbone which powers three core product areas for both B2C and B2B customers:
Proprietary trading - we use our technology to make trades in the sports markets on behalf of investors and ourselves
B2B models - we can supply data and predictive models to bookmakers to help them fine tune pricing and automate trades
StrataBet - our consumer insights platform is powered by our predictive models and used by bettors to show them the best betting opportunities in real-time, while they are watching the game
3. What are the most challenging aspects in your job?
At Stratagem, we are building and applying traditional quantitative finance models to the sports markets, which of course, are fundamentally different to the finance markets in many ways including their structure and liquidity. We are solving new problems with new technologies.
Doing something no one has ever tried before comes with lots of challenges, but that’s also one of the most interesting things about working here - that we are all empowered to work together and solve them.
4. For those who are already advanced in Machine Learning, what is Stratagem doing differently in this space that makes your company a particularly interesting place to work?
Stratagem is one of the few places I know that is tackling sports modelling and trading in a totally different way. Rather than employing an army of manual traders, we’re using Machine Learning to improve and automate signal extraction and execution.
In short, from the outside we may look like your typical systematic hedge fund but the underlying asset is a sporting event.
5. What checks and balances do you use to make sure that you don't make mistake
Testing and reviewing our work is essential in this industry. We are very vigilant in our code reviews and make sure we conduct the necessary reviews before changing our production code base. We are also very thorough in our unit testing and we test our code thoroughly in stress scenarios to see how our dynamic positions behave.
6. What are you doing to stay current with the latest technology?
It’s important to stay up-to-date with the latest news and research in the industry. I always put some time aside for reading publications and blog posts and books and recommend Berkeley’s BAIR blog.
I aim to visit around two conferences on AI and ML each year. In London, we are very lucky that there are many high-quality conferences and events where the data and science communities come together.
The AI and ML world moves fast, so at Stratagem we are also very focused on expanding our research network and sharing knowledge with our peers to keep our work current and make sure we are constantly advancing. We work with several leading universities including Liverpool University, Imperial College and UCL, undertaking research and funding important PhD projects for academics.
7. What data scientists do you admire? Which startups?
I have a lot of respect for the people that have inspired and educated throughout my career, including Trevor Hastie, Rob Tibshirani and Andrew Ng who are pioneers in the computer science and statistics space. Then of course the people that have led the AI revolution like Peter Norvig and Jeff Dean at Google as well as the team at DeepMind.
8. Is there anything you wish you had known before you began your career?
While the time I spent learning theoretical mathematics was captivating, I’ve been very much drawn to computational math’s. I enjoy implementing and thinking about practical algorithms a lot and wish I had of spent more time on this early on.
9. What is the best thing about working as a lead researcher and development engineer?
It’s my job to drive Stratagem’s research agenda forward in a way that is beneficial for the business and has a positive commercial impact. I enjoy working in this fast-paced world and managing multiple projects. We are combining cutting edge ML algorithms with very niche domain specific knowledge and our data scientists and football analysts need to work together to make this work.
10. Where do you see yourself in the future?
What we are working on for Stratagem is only just taking off. In the future, I see myself still here having established our company as a leader in the world of professional sports trading and continuing to push the boundaries of sports, and particularly, football modelling.