Built an open-source backtesting framework for prediction markets with chronological trade replay, performance metrics (Sharpe ratio, max drawdown, P&L by category) over 36GB of Parquet data, Kelly criterion position sizing, and fee modeling for Kalshi/Polymarket with longshot-fade and maker-replication strategies.

A full-stack investment platform enabling AI-driven portfolio optimization based on Modern Portfolio Theory with multi-source sentiment analysis using VADER and data from 5,000+ news articles and social media posts to generate actionable trade recommendations.

A tool that extracts intelligence from financial documents using Retrieval-Augmented Generation (RAG) to analyze 10-K, 10-Q, and earnings call transcripts, enabling interactive Q&A on complex financial reports and transforming unstructured filings into structured insights.

An AI-powered command-line tool that transforms terminal errors into clear explanations and actionable fix suggestions, featuring shell history integration.