Prize Winners

Monthly Prizes

Month 1

Prize winners chosen on 20th of January, 2020.

Congratulations to our Month 1 winners!

First place: Julian (username: julvo) – with a round total of 56.92, winning the monthly first place prize of £1,500.

“My background is in Computer Science (MSc, Imperial College) and Mechanical Engineering (BSc, RWTH Aachen). I worked in machine learning engineering roles as I enjoy both the modelling process as well as the deployment. Currently, I’m working for a small startup in London where my role is a combination of machine learning research and building the product. When developing models, first, I try to explore the data set to get an intuition for the data. Then, I’d focus on building a reliable evaluation system, to guide the modelling process and tell me whether I’m making any progress. Only then, I’d start developing the actual models. I’d start with super simple baseline models and gradually increase the model complexity to improve on the evaluation metrics. Generally, I try not to do too much feature engineering, but use end-to-end models where possible.”

Second place: Darin (username: darind) – with a round total of 56.53, winning the monthly second place prize of £750

“I’m a quantitative developer working in the finance industry, with experience at several large hedge funds. I believe the best model is the simplest model.”

Third place: Lucas (username: lum) – with a round total of 56.41, winning the monthly third place prize of £250

“I’m a research scientist working for Spotify in London. My work revolves around developing statistical models that match users with the right audio content at the right time. Before that, I did an PhD in machine learning at EPFL, in beautiful Switzerland. I like building predictive models of sports; together with a friend, I’m running https://kickoff.ai, a website that displays predictions for matches of the main European football leagues. I usually like to stick to simple models, especially in the case of non-stationary data and when the signal-to-noise ratio is low, such as for basketball. I’m a (pragmatic) Bayesian and I like to use Bayesian inference whenever possible.”

Month 2

Prize winners chosen on 17th of February, 2020.

Congratulations to our Month 2 winners!

First place: Jakob (username: jakkes) – with a round total of 61.29, winning the monthly first place prize of £1,500.

“I am currently at the end of my studies at KTH Royal Institute of Technology in Stockholm, Sweden, where I have been studying Engineering Physics (BSc) and Computer Science (MSc) focused on Autonomous systems. At the moment I am doing my master thesis at Ericsson and am expected to graduate in June. I really enjoy ML/AI, both applied and research focused – my favourite (sub)topic would be reinforcement learning and probabilistic graphical models. Much like Lucas, I am using a Bayesian approach to the modelling and did not do too much feature engineering. Furthermore, I also spent time at EPFL in (indeed beautiful) Lausanne, Switzerland, although only for one exchange semester during my master studies.

Second place: Lucas (username: lum) – with a round total of 60.94, winning the monthly second place prize of £750

Lucas won third place in Month 1, and his bio is above.

Third place: Steve (username: gasolboostingmachines) – with a round total of 60.83. 

“I’m an ex G-Research employee, I left at the start of 2019 to travel. I spent a fantastic 6 years at GR as a SysAdmin, DevOpser and then a Quant Analyst. This is the first data-competition I’ve done, I jumped in because I wanted to get into watching the NBA whilst I was travelling (it’s been hard to watch the football in some time-zones) and I knew it would be a decent comp as it’s run by GR.  I’ve spent most of my time reading NBA articles and working on engineering features that I feel have a tangible effect on the result of the game.  I’ve kept the modelling pretty simple until now, but I have been fiddling with some more complicated modelling structures that I’ll hopefully have some time to throw into the mix soon.”

As Steve is a G-Research affiliate, he is not eligible for our cash prize. Instead, Julian (username: julvo) who came in 4th with a round total of 60.71 has won our third place cash prize of £250. Julian came in first place in Month 1 as well, and his bio is available above.

Month 3

Prize winners chosen on 16th of March, 2020.

Congratulations to our Month 3 winners!

First place: Julien (username: julien-portier) – with a round total of 51.86, winning the monthly first place prize of £1,500.

“I am currently studying Mathematics at the Ecole Polytechnique in France, and I am about to intern in Millennium as an Algo Trader. I consider myself more of a mathematician than a computer scientist. I really enjoyed this competition and find it amazing that, beyond the math formulas, we can build models that are actually very accurate in the real world!”

Second place: Jakob (username: jakkes) – with a round total of 51.31, winning the monthly second place prize of £750

Jakob won first place in Month 2, and his bio is above.

Third place: Robert (username: lidergithub) – with a round total of 51.21, winning the monthly third place prize of £250

“I got a Ph.D. degree in electrical and computer engineering at the University of Florida. My dissertation is to develop algorithms and apply machine learning techniques to remote sensing data. I am a freelance data scientist and spend most of time on Kaggle.com to advance my skills in the real world data. My approach for this challenge is to get previous weekly accumulated features based on home and away games, including weekly Elo rating for pointsDiff, Blocks, Rebounds, Steals, and Assists. I also combine previous season results as additional features. I train these features by LightGBM and validate by the last 2 seasons to avoid overfitting.”