Junior Quantitative Researcher
About the role
A lot of what we trade starts as a hunch at the desk: a pattern that keeps showing up, a relationship that feels like it should hold, a setup someone wants to trust but doesn't yet. We're hiring a Junior Quantitative Researcher to turn those hunches into evidence, and evidence into signals the desk actually uses. You'll take questions from the traders, test them honestly against the data, and come back in plain language: what holds, what doesn't, how strong, since when, and where it breaks. This is not a systematic seat: your models won't trade on their own, and you won't be one factor in an anonymous alpha pool. Your work is done when a trader understands it, trusts it, and acts on it. It's an entry point into a serious seat: as your signals prove out, they get more weight in the book, and you get more scope. The role is full-time and fully remote, US-based, overlapping the desk on US market hours.
What you'll own
• The question queue. Desk hunches in, answers out. You scope the question, run the study, and come back sometimes same-day with what the data says, at the depth the question deserves.
• The signal shelf. A growing library of signals and screens the desk trades around (earnings and event setups, positioning and flow extremes, regime and factor context) each with a documented edge, known failure modes, and live tracking against expectation. You retire what decays.
• The evidence standard. Event studies, conditioning analyses, base rates, and walk-forward checks that hold up: point-in-time data, no leakage, no survivorship, costs counted, multiple-testing restraint. A backtest here is a way to kill an idea, not to sell it.
• The quant read. The recurring numbers layer under the desk's picture of the market (regime, factor moves, breadth, positioning) feeding the morning picture rather than duplicating it.
• The research data. Clean, point-in-time datasets (prices, fundamentals, earnings calendars, options and positioning) organized so every number you've ever reported traces back to code and data.
• AI leverage. Use modern AI tools to test more ideas, read more literature, and write cleaner code than one person otherwise could, and verify what they give you.
Who you are
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.
How you work
• 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.
Core skills
• 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.
What we offer
• A seat inside a live trading operation, with a direct line to the traders who act on your evidence.
• Fast feedback: when a signal proves out, you watch it get used - and you'll know precisely what your work changed.
• A well-capitalized firm with a distinctive approach to markets.
• A deliberate growth path: own the question queue first, then take on a bigger slice of the research agenda as you prove out.
• A small, low-ego, fully remote team.
• Compensation: $120k - $170k + bonus.
