Research Scientist, Machine Learning
Most ML research roles sit at some distance from a result that can be verified. The lab publishes. The paper gets cited. Whether the idea holds up under live conditions is someone else's problem. This role is structured differently.
We are a systematic trading firm where ML is not a support function. It is how the organization generates alpha. The research here is evaluated against financial markets, which are adversarial, non-stationary, and indifferent to elegant methodology. That is what makes the work serious, and what makes a genuine finding genuinely valuable.
We are looking for a researcher with a record of original contributions, deep command of modern ML, and the intellectual honesty to distinguish real results from artifacts. The problems are hard. The feedback is direct. The peer group is small and exceptional.
What you will work on
Research and modeling
- Develop and evaluate novel ML models applied to financial market data across the full research lifecycle, from hypothesis to production
- Design experiments with rigorous methodology; you are expected to be deeply skeptical of your own results
- Identify signal in high-dimensional, noisy, non-stationary environments where standard approaches fail
- Push the boundary of what is currently in production; original thinking is expected and rewarded here
Collaboration and culture
- Work on a small, senior research team where ideas are debated on their merits, not their provenance
- Present research clearly and defend methodology under rigorous peer scrutiny
- Uphold intellectual honesty and reproducibility as foundational standards, not aspirational ones
- Mentor junior researchers as the team scales
Breadth and frontier thinking
- Stay current with ML research and evaluate the applicability of emerging techniques to the firm's problem set
- Develop a personal research agenda, supported by real compute and colleagues worth arguing with
- Contribute to the firm's long-term thinking on where ML in financial markets is heading
- Work across varied data types, time horizons, and market regimes; no two problems look the same
Technical areas of focus
Deep learning Time series modeling Reinforcement learning Probabilistic modeling Representation learning NLP / alternative data Python PyTorch / JAX Distributed compute
Who you are
- PhD in machine learning, statistics, mathematics, physics, or computer science, or equivalent research output that speaks for itself
- A record of original ML research: published work, demonstrated impact in a prior role, or competition results with real substance behind them
- Deep command of modern ML methodology; you understand not just how to run models but why they work and precisely where they will fail
- Fluency in Python and the scientific computing stack; you write code that others can read, test, and build on
- Experience with, or genuine intellectual interest in, applying ML to financial markets, time series, or other high-noise sequential data domains
- An instinct for what makes a result real rather than an artifact; you are more interested in durable findings than impressive backtests
- Comfortable with autonomy in an environment with limited structure and high expectations
Why this role
The best ML researchers in quantitative finance cluster at a small number of organizations. This is one of them. The problems are real, the feedback is direct, and the culture treats research as a craft. You will not spend your time maintaining legacy systems or writing justifications for headcount. You will spend it thinking, building, and finding out whether your ideas hold up when tested against the one environment that does not care how good they look on paper.