Quant researcher vs quant trader vs quant developer: how the roles differ

Published 15 October 20267 min readStart learning quant:Part 1 Part 2Part 3Part 4 Part 5 Part 6 Part 7

A quantitative researcher develops and tests models or investment ideas. A quantitative trader uses quantitative reasoning to make and manage trading decisions. A quant developer builds the software and data systems that support research, pricing or trading. The responsibilities overlap, and titles alone do not tell you how a particular team works.

The most useful distinction is the output each role owns. Is your main deliverable evidence about a model, a decision involving positions, or software other people rely on? That question helps you compare jobs and choose a learning path.

This article compares role responsibilities and preparation. For examples of how the work unfolds, read what quants do all day. For the study sequence, use the quant roadmap.

Researchers, traders and developers share a data-to-decision workflow but take responsibility for different outputs: evidence, positions and reliable systems.

A comparison of responsibilities, not a universal organisational chart. Some teams combine these functions in one role.

The comparison at a glance

Dimension	Quant researcher	Quant trader	Quant developer
Main question	Does the model or hypothesis provide useful evidence?	What should we trade, quote or hedge, and under which limits?	Can the workflow run correctly and reliably?
Typical output	Experiment, model, forecast or research conclusion	Positions, execution decisions and managed exposure	Data pipeline, library, research tool or trading component
Useful foundations	Probability, statistics, research design and coding	Probability, market mechanics, risk and coding	Programming, algorithms, data and domain knowledge
An important failure	Leakage, weak validation or an unstable finding	Unmanaged exposure, misunderstood fills or poor execution	Incorrect data, unreliable operation or untraceable results
A first learning project	Test a simple hypothesis against a baseline	Analyse a specified order's execution cost	Package and test a numerical calculation

The table describes learning priorities, not a ranking. Researchers need engineering habits; traders need evidence; developers need to understand the decisions their software serves.

What does a quantitative researcher do?

A researcher turns a question into a testable model or experiment. The work may involve market data, economic observations, company information, pricing relationships or other inputs. The researcher chooses a representation, constructs a fair comparison, investigates failures and communicates what the result supports.

For example: does a proposed signal contain information beyond an existing baseline? You need to define the target, align observations with their availability times, separate development from evaluation, and account for costs where the intended use involves trading.

An improvement in a fitted metric is not automatically an improvement in a financial decision. A forecast can improve while turnover makes the strategy less useful. A model can appear stable until you change the universe or remove an accidentally forward-looking variable.

Citadel's research FAQ describes research supporting both fundamental investment decisions and systematic strategies. That is a reminder that research is broader than one style of automated trading.

Study first: Probability, Statistics and Financial Data, Econometrics and Financial Time Series, and Quantitative Research, Data and Backtesting.

Try this: take an original hypothesis, define a simple benchmark, and write down what result would make you reject the idea before you inspect the evaluation period. The time-ordered validation lesson gives the next layer of detail.

What does a quantitative trader do?

A quantitative trader connects information and models to trading decisions. Depending on the desk, the work can include quoting, monitoring automated strategies, managing inventory, adjusting risk, investigating unusual behaviour and reviewing executions with researchers and engineers.

Imagine a model estimates value at 100.20 while the market offers an opportunity to buy at 100.10. The difference of 0.10 is only the start of the decision. What size can you execute? How uncertain is the estimate? What happens to your existing exposure? Are fees, hedging costs and adverse selection large enough to consume the difference?

The trader's responsibility concerns the decision under those conditions. It does not imply that every trader manually sends every order. Automation can change the speed and mechanics while leaving meaningful decisions about risk and strategy behaviour.

Jane Street's interview guidance emphasises collaborative problem solving and says its trading interviews do not test prior finance or economics knowledge. That is useful evidence about its own process, not a rule for every quant trading employer.

Study first: Financial Instruments, Markets and Conventions, probability, and Market Microstructure and Execution.

Try this: calculate what happens when an order consumes several levels of a fixed order book. State the difference between the displayed midpoint, the average execution price and the cost after fees. Then read what market making is to connect fills with inventory.

What does a quant developer do?

A quant developer makes quantitative work usable through software. One role might focus on research tools and data pipelines; another on pricing libraries, execution infrastructure or performance-sensitive systems. The title does not identify the latency requirement or programming language by itself.

Suppose two researchers obtain different results from the same notebook. The engineering investigation may uncover different dependency versions, a silently updated dataset, an implicit timezone assumption or a model parameter stored only in memory.

Repairing that workflow requires more than speeding up the numerical calculation. You need explicit inputs, traceable versions, tests, failure handling and a repeatable execution process.

Research engineering often involves close collaboration with model users. A well-designed interface prevents an invalid unit or timestamp from quietly entering the experiment. A useful error message can save more time than a faster loop.

Study first: Python for Quantitative Finance, testing numerical code, and the domain programme your tool supports. Add Professional Quant Practice for reproducibility and documentation.

Try this: package a return calculator or pricer so another person can call it with documented inputs. Include tests for a simple benchmark and an invalid input. Use the Python beginner guide if the language is still new.

One project, three role-specific investigations

Take a hypothetical daily trading rule. Rather than creating three unrelated projects, investigate it through three lenses.

Research lens: define the hypothesis and evaluation design. Compare the signal with a simpler alternative, and check whether a timing repair removes the apparent advantage.

Trading lens: turn the proposed position into feasible orders. Examine spread, costs, inventory and the consequences of partial fills or an unavailable market.

Development lens: make the inputs and calculation reproducible. Detect duplicated timestamps, prevent look-ahead joins, and produce a decision ledger someone else can audit.

One illustrative trading idea viewed through research, trading and development: test the evidence, model feasible execution and make the result reproducible.

Different questions can use the same project. They require different evidence in the final report.

A good presentation explains which lens you investigated deeply and which you simplified. You do not need to imply that a beginner project reproduces an institution's complete production process.

Where do pricing, portfolio and risk quants fit?

The researcher–trader–developer comparison leaves out important specialisms.

A pricing quant investigates valuation, calibration, hedging and numerical implementation. A portfolio quant studies allocations, factors, constraints and investment risk. A risk quant develops or uses measurements of exposure and loss. A model-validation quant independently challenges a model's design, implementation and intended use.

These roles can also involve research and development. A pricing quant may write substantial software; a validation quant may build a challenger model. Start with the intended output and decision rather than forcing every vacancy into three labels.

For these routes, inspect Options Pricing, Portfolio Management and Quantitative Risk Management. The maths requirements article shows how the mathematical emphasis changes.

How to choose a role to investigate first

Ask which difficult activity you are willing to repeat. Do you enjoy investigating why an experiment failed? Making decisions with incomplete information? Tracing a software failure and improving a tool? Understanding a contract deeply enough to explain a pricing disagreement?

Use small projects to test those preferences. Enjoying a textbook subject does not tell you whether you will enjoy the associated workflow. Equally, one frustrating debugging session does not establish that engineering is the wrong path.

Then read vacancies and ask practical questions:

What output does the role own, and who uses it?
How does research become a production decision?
Which languages, data and systems are central to the actual work?
How are models reviewed and failures handled?
What does a junior person produce in their first months?

These questions reveal more than title prestige or an unsourced compensation comparison. This article deliberately does not present a salary league table: geography, experience, firm, performance and pay structure can dominate title-level comparisons.

Frequently asked questions

Which quant role requires the most maths?

There is no reliable universal ordering. Research and pricing roles can demand substantial probability, statistics or stochastic analysis; engineering roles can demand deep numerical or algorithmic reasoning. Examine the particular work and its mathematical dependencies.

Can a developer move into research or trading?

Skills can transfer, especially domain knowledge and reliable implementation, but the destination role adds responsibilities that require evidence. A transfer is not automatic. Test the missing research or decision-making skills in a scoped project and compare them with actual role requirements.

Does every quant developer need C++?

No single language requirement covers the whole category. Some roles emphasise C++ and systems performance; others focus on research infrastructure or scientific Python. Learn Python to begin numerical work, then choose additional skills from the role and project requirements.

Choose the work, then the preparation

Pick one role family and complete the associated first project. Use the SelfMadeQuant roadmap to connect it to programmes, then use interview preparation to practise explaining the concepts behind your decisions. A role choice becomes more informed when it rests on work you have actually tried.

Download the accompanying Python examples

Get the series code package. The download contains four runnable beginner examples, an editable notebook and generators for the compounding, portfolio-correlation and Monte Carlo charts. Start with the unchanged examples, then alter one assumption and explain the result. Downloads use the app’s usual email or member access.