Technical Projects

A local-first desktop application for systematic vocabulary acquisition through contextual reading. Built on a 97,000-entry lexical database with real-time NLP tokenization, spaced-repetition tracking, and an embedded collection of 21,000+ curated articles.

Built to solve the creator's own analysis paralysis with color selection—demonstrating tools born from real creative need. Analyzes artwork and recommends specific colored pencils using the Delta-E 2000 algorithm for perceptually accurate color matching. Deployed to production in 2 weeks.

Interactive automation tool that removes decision fatigue from complex multi-step workflows through intelligent task orchestration. By visualizing process dependencies and handling routine decisions, it frees users to concentrate on strategic thinking and creative problem-solving.

Explores how local LLMs can personalize education by mapping individual desires, behaviors, and passions to learning outcomes. Deployed on Raspberry Pi 5, the system uses multi-modal interaction to adapt to each learner's unique context—without cloud connectivity or data extraction.

Deep Convolutional GAN for generating album cover artwork, trained on 20,000 diverse album covers. Implements advanced DCGAN stability techniques to produce colorful, diverse 128×128 RGB designs.

AI-powered discovery system that analyzes the GitHub ecosystem to find open source projects aligned with career goals and learning objectives. Transforms repository browsing into strategic opportunity analysis.