Back to Home

Projects

A collection of my work in quantitative finance, machine learning, and systems programming

portfolio_optimizer

portfolio_optimizer

A comprehensive portfolio optimization tool implementing Modern Portfolio Theory to help investors make data-driven allocation decisions based on historical performance data.

Key Features:

  • Multiple allocation strategies (equal weight, minimum variance, maximum Sharpe ratio)
  • Configurable rebalancing periods
  • Excel export functionality for analysis
PythonPandasNumPy
trends

trends

A machine learning model that classifies market regimes with impressive out-of-sample accuracy, helping traders identify trend directions and durations.

Key Features:

  • 71% out-of-sample accuracy
  • ±2 day duration accuracy
  • 10-17 day average trend predictions
PythonPyTorchTransformers
trading_strat

trading_strat

A sophisticated backtesting framework designed for systematic trading strategy development, featuring comprehensive visualization tools and performance analytics.

Key Features:

  • Interactive HTML visualizations
  • Comprehensive equity curve analysis
  • Multiple strategy support and comparison
PythonPandas
chess_engine

chess_engine

A modern chess engine implementation combining Rust's performance with OpenGL graphics for smooth, real-time gameplay visualization.

Key Features:

  • Complete chess rules implementation
  • Move validation and legal move generation
  • OpenGL-based rendering for smooth graphics
RustOpenGL
dwl

dwl

A personalized fork of the dwl Wayland compositor, featuring custom patches and configurations for an optimized Linux desktop environment.

Key Features:

  • Custom patches and modifications
  • Lightweight and performant
  • 2 GitHub stars
C