Skip to content
ALL WORK
MyMedicalHUB2024 — 2025iOSAI & ML

Atlas Athlete

AI fitness assistant using on-device pose estimation and NLP injury-risk classification for personalised training.

Atlas Athlete takes the same musculoskeletal assessment engine behind EMMA and points it at athletic performance rather than clinical rehabilitation. A full-body movement screen runs from the phone — no wearables, no sensors — and produces a picture of asymmetry, imbalance and movement inefficiency.

My work covered the on-device pose estimation pipeline and the NLP models used for injury-risk classification, plus the real-time movement feedback that turns a raw pose stream into something an athlete can act on mid-set.

Highlights

  • Full-body screen covering 350+ joints and motion patterns from a phone camera.
  • Injury-risk classification driving personalised exercise recommendation.
  • Real-time movement feedback during exercise.

Gallery

Three Atlas Athlete screens — the athlete dashboard, the Ask EMMA assistant, and a movement assessment report
Dashboard, assistant and report