quantlab

A personal research repository exploring quantitative finance, model risk, and machine learning for derivatives pricing and hedging.

🔍 Note: This project is unrelated to QuantLib.org. It is a private learning and experimentation space by AlmostAnna.

This repo contains:

• Classical models: Black-Scholes, Heston
• Hedging analysis: Error decomposition, discrete trading
• Machine learning: Buehler-style deep hedging with transaction costs
• Stress testing: Sensitivity to volatility misspecification, rebalancing frequency

All code is organized to support reuse, clarity, and diagnostics—not just one-off experiments.

quantlab/ 
├── notebooks/ 
│ ├── models/ # Stochastic volatility, etc. 
│ ├── model_risk/ # Hedging errors, replication failure, Greeks 
│ └── ml/ # Deep hedging, training diagnostics 
├── src/ # Reusable quant primitives (installable as 'quantlab')
| ├── quantlab/
│ ├── calibration/ 
| ├── data/ # Synthetic data generation
│ ├── hedging/ # Greeks, naive strategies
| ├── instruments/ 
│ ├── market_data/ 
| ├── ml/ # Models, metrics
| ├── models/
| ├── pricing/
| ├── sim/ # MC simulations
│ └── utils/ 
├── ml/ # ML-specific training and evaluation
├── tests/ # Tests
├── pyproject.toml # For editable install 
└── environment.yml

1. Clone and install:

   git clone https://github.com/AlmostAnna/quantlab.git
   cd quantlab
   pip install -e .[dev]

• Clarity over cleverness: Code should speak for itself.
• Model risk matters: Every assumption is surfaced and tested.
• ML as a tool, not a black box: Diagnostics, baselines, and stress tests are first-class citizens.

© 2025 - 2026 AlmostAnna — For learning, reflection, and professional growth.