Hedge Fund

What a Quant Actually Does Is Mostly Data Work

The job is imagined as inventing mathematics. In practice a large share of it is cleaning data, building infrastructure, and testing ideas that turn out not to work.

↩ Looking BackPart of the 2020 to 2026 retrospective, written in July 2026. The date below marks the 2024 events this piece revisits, not when it was published, so it draws on everything known through mid 2026.
Nathan Xiang·March 9, 2024

The Categories

Quant is not one job. It covers several roles with different daily work and different skills.

RoleFocus
Quantitative researcherFinding and testing signals or strategies
Quantitative developerBuilding the systems that run them
Desk quantPricing and hedging models for a trading desk
Risk quantMeasuring and modelling exposure

The distinction between researcher and developer is frequently blurred, and firms increasingly expect both. A researcher who cannot implement their own ideas is slower than one who can.

Where the Time Goes

The consistent report from people doing the job is that a large majority of the work is data.

Acquiring it, cleaning it, aligning timestamps, handling corporate actions, dealing with revisions, identifying errors, and building the infrastructure that makes it queryable. Only after all of that does anything resembling research begin.

The interesting part of the job depends entirely on the boring part being done correctly, and the boring part is most of the hours.

The reason it matters so much is that data errors produce false results that look exactly like discoveries. A survivorship biased universe, a mistimed data field, or an unadjusted stock split will generate an apparently profitable strategy, and nothing about the output distinguishes it from a real one.

The Failure Rate

Most hypotheses do not work. A researcher may test many ideas for each one that survives to production, and the ones that survive frequently deliver less than the research suggested.

This is a difficult working environment psychologically. The work is intellectually demanding, the failure rate is high, and the feedback loop on whether a deployed strategy is genuinely working takes a long time because of the sample size problem.

Firms manage this by measuring process rather than outcomes over short horizons, which is easier to state than to implement.

What the Mathematics Actually Involves

The background expected is genuine: probability, statistics, linear algebra, optimisation, and for derivatives roles stochastic calculus.

What is used daily is more prosaic. Regression, hypothesis testing, and careful handling of sample size occupy far more time than advanced techniques. Machine learning methods are widely used and the harder problem is nearly always preventing them from fitting noise rather than applying them.

Programming ability is not optional. Python dominates research, with C++ used where latency matters. Comfort with databases and version control is assumed.

Where the Roles Sit

Hedge funds and proprietary trading firms hire researchers to find strategies, with compensation frequently tied to the performance of what they build.

Banks hire desk quants to price and hedge products, and risk quants to model exposure and meet regulatory requirements. The work is more structured and the compensation less variable.

Asset managers hire quants for portfolio construction, factor research, and risk systems, sitting somewhere between the two in both style and pay.

What Distinguishes People Who Are Good At It

Scepticism about their own results, which is the single most important trait given how easily false discoveries appear.

Understanding of markets rather than only of mathematics, since a statistically valid relationship with no economic mechanism is usually noise.

Engineering discipline, because research that cannot be reproduced or deployed reliably is not useful.

And tolerance for the failure rate, which is the thing that most often causes people to leave.

The Bottom Line

Quant roles split across research, development, desk pricing, and risk, and all of them involve far more data work than the title suggests. The failure rate on hypotheses is high, data errors reliably produce results that look like discoveries, and the most valuable trait is scepticism about your own findings. The mathematics is real and the daily bottleneck is usually the data.

Explore Teen Biz News →