/projects

FitOntology

a client-intelligence layer for personal trainers

  • Python
  • TypeScript
  • Next.js
  • DuckDB

fitontology.conalhickey.com ↗

FitOntology unifies a personal trainer’s scattered inputs — wearables, client intake, and ACSM training guidelines — into one per-client readout: a recommendation for each client with its rationale, the flags that triggered it, and the trend behind them. It surfaces a roster, an action queue (“who needs attention today”), and a calibration view that learns from which recommendations a trainer accepts or overrides.

The analysis is rules-based and conservative — it cites the evidence behind every call and refuses to overreach. The demo runs on five synthetic personas, read-only. Source is on GitHub; it’s the trainer-intelligence project AthDash grew out of.