At Altas Technologies we build robust, fully automated systems to predict and trade cryptocurrency markets and are in the process of extending our trading capabilities from crypto to global equity markets. You’ll join at a very exciting time in the growth of the team, and will be able to contribute to our success in scaling the strategy beyond crypto to the equities markets. As a technology-oriented and scientifically-minded group, we design and build our own cutting-edge systems, from high-performance trading platforms to large-scale data analysis and research infrastructure.
Your Core Responsibilities
As a Quantitative Researcher, you will be operating as a senior and core member of our high-performance, dedicated, multi-disciplinary team, with a background in Mathematics, Physics, Computer Science, Robotics, Data Science, Engineering, Statistics. The one thing we all have in common, regardless of our individual backgrounds, is a creative and scientific mindset.
You will not get a lot of top-down instruction telling you exactly what to do and when to do it. You'll get directional advice, useful frameworks, access to interesting challenges, and plenty of freedom to execute as you see fit. We are meritocratic by nature and believe that empowering talent in our organization is the only way to achieve our highly ambitious goals.
We are looking for a creative scientifically minded individual who can lead the growth of the firm’s quant research to improve prediction quality in delta 1 mid-frequency (intraday) trading context. You’ll make a real impact on our alpha generation, feature modelling, machine learning ensembles, alternative data research, and R&D quality standards generally.
- You will work together with other quants on alpha generation, feature modelling, portfolio optimization, and extracting predictive signals from public feed and alternative data more generally. More specifically, developing orthogonal alphas across multiple categories.
- Develop and improve sampling-, weighting-, cost-, target-, hyperparameters- and (network) architecture functions for model trainin
- Develop ML/network architectures (MLP, CNN, GNN) for custom prediction problems in trading, including data representations
- An advanced degree with excellent grades in Mathematics, Physics, Computer Science, Data Science, Engineering, Statistics or any other highly quantitative field
- Experience as Researcher in (mid-frequency / intraday) D1 predictive modelling, developing orthogonal features across multiple categories in statistical arbitrage context.
- Significant practical experience with at least one mainstream ML (XGB, (C/R)NN, LSTM, RF, …) approach to improve prediction quality in a trading context, being aware of all ways one can overfit in the process. Solid knowledge of probability, (Bayesian) statistics, (non-convex) optimization, and mathematical modeling more generally
- Experience extracting predictive signals (alphas) from both public feed and alternative data.
- Scientific mindset with experience in numerical programming with Python. Detail oriented and highly structured, using a rigorous process to ensure your results are reliable and well tested
- Serious about clean code, simple but well-architected systems, and continuous improvement
- Sense of ownership taking full responsibility and accountability for your contributions
- Technical accomplishments are considered a plus: Kaggle, Hackathons, Olympiads, academic publications in e.g. NeurIPS, ICML.