Clap evolved from dataset plumbing into a private-first operator system with simulation-first execution and local training paths.

QERS moved from architecture pitch mode to measurable profile-based simulator evaluation with practical gap metrics.

A practical update on tiered seed-band modeling, routing logic, and assignment optimization for NCAA seed prediction.

A weekly live update covering technical progress, project movement, friction, and next focus.

LOL News started as a playful React Native app idea. The useful work was in feed reliability, voice output behavior, and content filtering choices.

Convo AI started as a gamified learning prototype. The hard part became progression logic, state consistency, and writing tasks that do not feel fake.

AI Ethics Coach is a Chrome MV3 extension that reviews selected text and returns practical flags without pretending to replace policy or judgment.

This book project started as a prompt pattern guide and turned into a stricter exercise in naming, scope control, and evidence-backed examples.

OpsPilot is a local-first workflow agent project. The hard part is contracts, guardrails, and testable execution paths.

Notes from the Final Four Analytics Challenge—Python, Zerve, public leaderboard, and what happens when March ignores your model.

A compressed build on a telehealth-flavored portal—knowledge graph, Twilio, Claude, and the gap between demo and product.

Build notes from HPC research on BigRed200—evidence bundles, reuse distance, and why the first few pipelines did not count.

The Spark It started as a class group project and then led to a question: Could a board game learn how you play and respond proactively? Between my long interest in mechanical design and a general knowledge of neural networks, the idea found my credibility in my analysis. I thought 'EASY' - just merge tactile gameplay with adaptive intelligence. Instead of static rules, what if each match evolved, learning from how players strategize, take risks, and bluff? Similar to a digit