Start with a question you can measure
Quantitative finance brings together several kinds of work. A credit modeller, an options researcher and an execution engineer share mathematical tools, but they solve different problems. Choose a small project, then learn enough theory and engineering to complete it carefully.
Understand the objects
Read probability and statistics, market microstructure and data quality. Calculate returns and explain which information is available at a decision time.
Build a reproducible experiment
Use the research workflow and workbench. Define a target, create trailing features, fit a baseline and preserve an untouched evaluation period.
Turn a prediction into a decision
Study portfolio and risk constraints and execution and accounting. Include the costs that separate a useful forecast from a useful strategy.
Practise, then explain the result
Use Tomorrow Markets or an edition-compatible competition backtester. Write down what worked, what failed and which assumptions limit the result.
Explore a specialism
| Path | Main questions |
|---|---|
| Banking & credit risk | Default, loss severity, exposure and model validation. |
| Actuarial science | Claims, survival, reserving and long-term liabilities. |
| Algorithmic trading | Signals, execution, inventory and market structure. |
| Systematic investing | Allocation, factors, turnover and portfolio construction. |
| Quantamental investing | Economic hypotheses, point-in-time fundamentals and valuation. |
| Risk management | Exposure, tail outcomes, scenarios and operational controls. |
What counts as progress?
A complete small project is more useful than an impressive collection of disconnected models. Aim to produce a data description, a reproducible script, a baseline, a fair comparison and a short account of limitations. The repository map connects the existing collections to those tasks.