We are hiring a Machine Learning Researcher to build models that extract signal from noisy, non-stationary market data without relying on leakage or overfit. The role is suited to someone who can move between theory, empirical testing, and production constraints without losing skepticism.
You will work on representation, prediction, validation, and model diagnostics across market datasets. The goal is not to add complexity for its own sake. The goal is to find structure that remains useful after costs, regime shifts, and portfolio interaction effects.
What You Will Do
- Build machine learning models for noisy, high-dimensional financial data.
- Design validation protocols that control for leakage, overfitting, and unstable correlations.
- Study feature stability, regime sensitivity, and model degradation.
- Work with researchers and engineers to move approved models toward production.
- Document assumptions clearly enough for review, reproduction, and retirement.
What We Look For
- Strong machine learning, statistics, or applied mathematics background.
- Experience with Python and modern ML tooling.
- Comfort with non-stationary data and adversarial validation.
- Good taste in model complexity and diagnostic evidence.
- Interest in markets, but no need for a traditional finance background.
How Success Looks
Success means models become more useful, not merely more complex. The strongest work will improve signal quality while making uncertainty, failure modes, and live degradation easier to see.