We're looking for people with a strong ML background to work across Quant Research and Feature Engineering. No finance experience required - we'll teach you everything, from market microstructure to portfolio construction. We care about how you think and your ability to ask the right questions of data.
What you'll be doing:
- Formulate and test hypotheses about market inefficiencies - and stay honest when the backtest says no.
- Think ahead about how model predictions will be integrated into trading - directional bets, spread convergence, or something else.
- Design features with genuine predictive power, from classical statistical transformations to learned representations. Every feature is a hypothesis about the market encoded in a number, and you own the quality of that input.
- Work with alternative data, time series, and nonlinear dependencies to find signal in places others overlook.
- Build data pipelines that work reliably in production, not just in a notebook.
What we're looking for:
- Deep ML understanding, not just knowing the APIs, but having a clear sense of why gradient boosting tends to outperform transformers on tabular data, when a Bayesian approach beats a frequentist one, and what information leakage looks like in a time series context.
- Proven hands-on experience with both deep learning and gradient boosting frameworks – a strong conceptual understanding of neural network internals is essential.
- Serious attention to data leakage. In finance, the future bleeds into the past in non-obvious ways, and you need to spot these issues before they invalidate a backtest.
- Solid math background: statistics, probability theory, stochastic processes.
- Solid Python skills. Experience with pandas, polars, or streaming data processing is a plus.
- A statistical mindset: you don't just apply methods, you understand their assumptions, limitations, and failure modes.
- Professional English proficiency at B2 level or above.
Nice to have:
- Graduate of SHAD or a similar top-tier technical or quantitative program.
- Background in competitive mathematics, physics, or computer science olympiads.
- Demonstrated performance in Kaggle or other ML competitions.
