Jane Street
In-House Tools
The proprietary tools and systems that Jane Street built internally to power their quant trading operations — from Figgie to Iron and beyond.
Dune: The OCaml Build System
Jane Street created Dune (formerly Jbuilder), now the standard build system for OCaml projects worldwide. It handles compilation, linking, testing, and dependency management for OCaml codebases.
Key Features
- Incremental builds: Only recompiles changed files.
- Automatic dependency detection.
- Cross-platform: Works on Linux, macOS, Windows.
- Multi-package support: Manages monorepos with hundreds of packages.
Why It Matters
Dune is now used by thousands of OCaml projects globally, demonstrating Jane Street's influence on the broader OCaml ecosystem.
Dune documentation: https://dune.readthedocs.io/
Jane Street's Internal Trading Infrastructure
Jane Street has built extensive proprietary infrastructure for their trading operations:
Key Systems
- Execution Management System (EMS): Routes orders to exchanges globally.
- Risk System: Real-time position and risk monitoring across all books.
- Market Data Infrastructure: Processes millions of market data events per second.
- Research Platform: Backtesting and simulation environment.
Design Philosophy
- Correctness First: OCaml's type system enforces invariants at compile time.
- Observability: Every system is instrumented with metrics, logs, and traces.
- Resilience: Systems designed for graceful degradation during market stress.
Reference: https://www.janestreet.com/technology/
Figgie: The Trading Game
Figgie is a trading game invented by Jane Street for evaluating trading intuition and market-making skills. It's played in 4 rounds with 5 players and uses a unique card deck mechanics.
Game Mechanics
- Each player receives 10 cards from a standard deck.
- A "goal suit" is secretly selected; cards in that suit are worth money.
- Players trade cards, trying to accumulate goal suit cards.
- Trading requires quoting bid/ask prices and making markets.
Why Jane Street Uses It
- Evaluates: Market-making intuition, risk assessment, decision under uncertainty.
- Tests: Ability to update beliefs from new information (Bayesian reasoning).
- Used in: Recruiting events, internship programs, onboarding.
Play Figgie: https://www.figgie.com/
Iron: Code Review and Continuous Integration
Iron was Jane Street's internal code review and release management tool, which they used for over a decade before open-sourcing parts of it.
Iron Features
- Incremental reviews: Reviewers see only changed hunks, not full diffs.
- Release queues: Manages the process of releasing code to production.
- Integration with Dune for build and test automation.
Open Source Contributions
Jane Street's philosophy is to contribute back to the open-source community. They maintain dozens of libraries on GitHub:
- Core, Async, Incremental, Bignum, Bin_prot, and more.
GitHub: https://github.com/janestreet
Research and Simulation Platforms
Jane Street's research infrastructure enables rigorous quantitative research:
Key Components
- Historical Data Store: Tick-level historical data for all traded instruments.
- Backtesting Engine: High-fidelity simulation of trading strategies.
- Statistical Analysis Tools: Custom statistical testing frameworks.
- Visualization: Interactive research dashboards for PnL attribution.
Research Process
- Hypothesis generation from market observation.
- Statistical validation on historical data.
- Paper trading in a simulated environment.
- Gradual live deployment with risk controls.
Learning Resources
- QuantLib: Open-source quantitative finance: https://www.quantlib.org/
- Zipline (Quantopian's backtester): https://zipline.io/
Further Reading