Mining Digital Work Artifacts
AI-powered desktop app that mines developers' local work artifacts to generate portfolios and surface skill growth.
Overview
COSC 499 capstone project built with a team of 6. A cross-platform (Windows/macOS) desktop app that parses Git history, codebases (Python/Java/C++), documents (PDF/DOCX), and media to extract metadata, LOC, functions/classes, and per-contributor metrics. Integrated Google Gemini Pro for long-context semantic code reviews covering architecture, complexity, and security — with consent gates and audit logs for privacy-first compliance.
Key Highlights
- Team of 6 — led backend API design and Git parsing pipeline
- Parses Git commits, Python/Java/C++ codebases, PDFs, and DOCX for metadata extraction
- Google Gemini Pro integration for semantic code review (architecture, complexity, security, skill maturity)
- Electron + React dashboard with skills timelines, activity heatmaps, and top-project showcases
- One-page resume generator with PDF/HTML export
- Privacy-first: explicit consent gates, audit logs, and secure deletion receipts
More Projects
F1 Intelligence Hub
Production-grade Formula 1 analytics platform with real-time ML predictions and race strategy insights.
Jersey Number Recognition Pipeline
Spatio-temporal pipeline for recognizing jersey numbers from soccer player tracklets, combining pose-guided torso cropping with CNN and BiLSTM recognition models.