Featured project
G.R.I.T.
A hypertrophy training product built around practical lifting, deterministic programming rules, workout logging, and careful AI boundaries. It is where fitness, product craft, and backend discipline meet.
Product type
Training PWA + native app
Core stack
Next.js · TypeScript · Supabase
Engineering theme
Rules engine + trusted AI boundary
The problem
Most workout apps either log too little or decide too much.
G.R.I.T. is built for the middle path: a focused hypertrophy workflow that helps lifters plan, log, and progress without burying them under clutter. The important training decisions stay explicit, tested, and reviewable.
Rules first, AI second
Training science decisions live in deterministic code and tests. AI can suggest exercise choices and explanations, but it does not own sets, reps, progression, or safety boundaries.
Native parity matters
The PWA is being shaped to feel like the native app, not a generic web dashboard. Useful web advantages are welcome, visual drift is not.
Training data should teach
Workout logs, PRs, soreness, and volume are most valuable when they turn into understandable feedback instead of raw numbers alone.
Future trainer, trusted boundary
A future AI trainer should explain, coach, and propose, then wait for confirmation before changing anything important.