1 Who this is for
Engineers who can already build things — and want to build them in markets.
The course assumes you are comfortable with Python, SQL, version control, basic statistics and the ideas behind machine learning. It assumes nothing about finance: every instrument, every convention and every piece of jargon is introduced from zero. The math is taught intuition-first; if you once passed a calculus and a linear-algebra course, you have enough to start.
| You bring | The course adds | Where |
|---|---|---|
| Software engineering, testing, CI | Backtests designed so they cannot lie; live systems with kill switches | Units 6, 15 |
| Data engineering | Point-in-time data, corporate actions, survivorship, market data at tick level | Unit 4 |
| ML / AI | Finance-specific validation, labeling and leakage; where ML genuinely helps | Units 12, 13 |
| Statistics basics | Fat tails, volatility, inference under multiple testing, the math of overfitting | Units 3, 5, 6 |
| — | Markets, instruments, derivatives, portfolio construction, risk, execution, careers | Units 1–2, 7–11, 14, 16 |
2 How long it takes
About three hundred hours, end to end — one hundred sessions of roughly an hour and a half, plus the labs and homework.
| Pace | Hours / week | Spine only (U1–U6, U9) | Whole course |
|---|---|---|---|
| Evenings & weekends | ~8 | ~4 months | ~9 months |
| Serious part-time | ~15 | ~2 months | ~5 months |
| Full-time sprint | ~35 | ~3 weeks | ~2 months |
The spine (Units 1–6 and 9) is what every quant role shares. The track matrix on the home page shows which of the remaining units matter most for researchers, traders, quant developers and independent traders.
3 Set up the labs
Two ways to run a lab: one click in Colab, or locally in a reproducible environment. Session 1.5 walks through the local setup in depth.
Fastest: open in Colab
Every session page has a 📓 Companion lab strip under its title. Open in Colab runs the notebook in your browser; its first cell installs the course package.
Local: clone and sync
Install uv, then clone the repository and create the environment — uv reads the lockfile, so you get exactly the versions the labs were tested with.
Run a lab
Open any
labs/uNN/sSS_*.ipynbin Jupyter or VS Code, or run its.pytwin from the terminal. Labs are seeded: you get the same numbers the page quotes.
$ git clone https://github.com/georgevrs/quant-notebook $ cd quant-notebook $ uv sync # core labs; add --all-extras for derivatives, ML, live trading $ uv run python labs/u02/s03_returns_pnl_leverage.py
Most financial data cannot legally be redistributed, and free sources change or disappear. So labs default to seeded synthetic data generated to have the properties of real markets (fat tails, volatility clustering, factor structure). It has a second, bigger advantage: with synthetic data you know the truth, so you can test whether a method finds a real signal — and whether it "finds" one that isn't there. Labs that need real data fetch it from free sources into a local cache.
4 How to read a page
Colour means the same thing everywhere — in callouts, diagrams, badges and charts — so you learn to read it once.
| Callout | What it holds |
|---|---|
| 📌 Key idea | The takeaway of a section — if you remember one thing, remember this. |
| 💡 Tip / Looking ahead | A shortcut, or where this idea returns later in the course. |
| ⚠️ Pitfall | The common mistake — usually the one that costs money. |
| 🔧 Try it / Homework / Lab | Hands-on work with a time estimate and a checkable result. |
| 🧭 Margin note | A clarifying aside or a rule of thumb. |
Formulas are real MathML rendered by your browser — no JavaScript, and they work offline. Numbers in the text that come from a lab are generated from the lab's output, so they always match what you get when you run it. Diagrams are drawn by hand; charts are drawn from lab data.
5 Errata & contributing
Finance moves and mistakes happen. If something is wrong, out of date or unclear, say so.
- Report an erratum — session, section, what is wrong, what it should say.
- The repository — every page, lab and tool is open source. Read
CONTRIBUTING.mdbefore a pull request. - Licences: text and diagrams CC BY 4.0; code (labs,
quantnb, tools, and code blocks in pages) MIT. Reuse it, teach with it, credit it.
6 Disclaimer
Read this once, and take it seriously.
Quant Notebook is educational material only. Nothing in it is investment, financial, legal, tax or accounting advice, or a recommendation to buy, sell or hold any instrument or to follow any strategy. The authors are not registered investment advisers, broker-dealers or financial planners.
Backtests, simulations and examples are hypothetical: they are built with hindsight, ignore or approximate real-world frictions, and do not represent actual trading. Past performance — real or simulated — does not predict future results. Trading and investing involve substantial risk, including the loss of more than your initial capital when leverage or derivatives are used.
Laws, tax rules and market regulations differ by jurisdiction and change over time; any regulatory detail in the course is a dated example, not guidance for your situation. You are solely responsible for your own decisions; consult a qualified professional before acting. The material is provided "as is", without warranty of any kind.
Educational material only — not investment, legal or tax advice. Backtests and simulations are hypothetical. Disclaimer · Text CC BY 4.0 · Code MIT.