We hire for demonstrated judgment and how you think, not for pedigree. This is a junior seat, so we don’t expect a markets résumé — the best evidence usually comes from wherever you’ve already done rigorous quantitative work. We look for signs that you are:
• Rigorous where it counts. You think in base rates, sample sizes, and conditioning; you know what autocorrelation, overlapping windows, and fat tails do to naive statistics, and you’d rather report a small honest edge than a large fragile one.
• Hypothesis-driven. You start from a mechanism (why would this work, who’s on the other side, why hasn’t it been arbitraged away) and let the data disappoint you, not the other way around.
• A translator. Your finished product is a chart and a paragraph a non-quant acts on. If the desk can’t understand it, it isn’t done.
• Comfortable finding nothing. “There’s nothing there” is a result you deliver without flinching: a clean negative saves the desk real money, and you never dress noise up as signal to give someone the answer they wanted.
• Genuinely curious about markets. You want to know why prices move. You don’t need professional markets experience, you need to care about the answer.
• Low ego and coachable. You take feedback well, update quickly when the facts change, and care more about the answer than the credit.
• At desk speed, without cutting corners. A rough answer today often beats a perfect answer next week; a deep study is worth a month when the stakes justify it. You know which question is which, and you label your answers accordingly.
• Kill your own results first. Before anyone else sees a number, you’ve gone hunting for the leak, the regime dependence, and the artifact that would explain it away.
• Reproducible by default. Versioned data and code; any signal or study can be re-run months later and give the same answer.
• Signal over noise. You surface the few things that matter, track what you’ve shipped, and say so plainly when something stops working.
• AI-native. Fluent with modern AI tools for research, coding, and literature triage, and disciplined about checking their output.
• Self-directed. You thrive working remotely with low guardrails, managing your own time and flagging what needs attention without being asked.
• Statistics and probability. Regression and its failure modes, hypothesis testing, bootstrap and resampling, thinking clearly about uncertainty in small and messy samples.
• Python data stack. pandas, NumPy, SQL, and plotting that makes a point; notebooks that read top to bottom, graduating to scripts when a study becomes a signal.
• Backtest and event-study hygiene. Point-in-time discipline, survivorship and look-ahead awareness, transaction-cost sanity, walk-forward validation, restraint about how many things you tested.
• Market data. Comfort with prices, returns, fundamentals, and earnings calendars; options or positioning data a plus.
• ML as a tool, not an identity. Regularized regression and gradient boosting when they beat something simpler, interpretability first. This is a statistics-first seat, not a deep-learning one.
• Crisp communication. Compressing a study into exactly what a trader needs to know, now.