Pozify won 1st place in the Backyard AI — Track at the Hugging Face Build Small Hackathon, earning a $4,000 prize. The official results are published on the Build Small prize ledger.
Backyard AI is the hackathon’s flagship track—described by organizers as the home for the most polished, useful builds. For Pozify to land first place there is a strong signal: small models can win when the product is specific, inspectable, and grounded in real user needs.
What Pozify Does
Pozify is a small-model workout form reviewer. Upload a short exercise video and Pozify returns a structured coaching report: exercise detection, rep counting, per-rep issue markers, annotated video output, confidence notes, and a grounded coach summary.
Video as evidence
Pozify starts with a real workout clip, then extracts pose and movement structure before writing any coaching text.
Rep-level feedback
The report separates exercise detection, rep counting, valid variations, and issue markers instead of returning generic advice.
Small-model pipeline
MediaPipe pose extraction, a tiny trained router, exercise rules, knowledge cards, and a small summary model—each with a clear job.
Grounded summaries
A verifier checks that the coach summary stays tied to the structured evidence and avoids unsupported safety claims.
Why Backyard AI
The Build Small constraint was the point: build something useful without hiding behind one giant opaque model.
Pozify fits that philosophy by keeping the system modular and inspectable. Pose extraction handles body landmarks. A custom PyTorch BiLSTM routes clips into squat, push_up, shoulder_press, or unknown. Exercise-specific logic counts reps and marks issues. A coach-summary model turns the structured artifacts into readable feedback.
Structured evidence comes first; language comes second. That design choice—specific models for specific jobs, with deterministic verification on the output—is what makes Pozify feel like a product rather than a chatbot wrapper.
What This Means
The $4,000 prize validates the direction we set out to prove with Pozify: at-home training feedback can be affordable, private, and specific enough to act on—without pretending to be a medical device or a full-time personal trainer.
We are grateful to Hugging Face and the Build Small organizers for running a hackathon that rewards useful, inspectable small-model systems. Winning the flagship Backyard AI track is encouragement to keep pushing on broader exercise coverage, stronger quality checks, and more independent evaluation.
Explore Pozify
Pozify is open source and available as a Hugging Face Space. Try the app, inspect the pipeline, and follow the small-model build.
Learn More
For the full product breakdown—the pipeline, model stack, coach intelligence layer, and trust boundaries—see our earlier post introducing Pozify for the Build Small Hackathon.
Introducing Pozify → · Official prize ledger →
The Team
Tien Ngoc VietYour coach, without the crowd.
Thank you for supporting Tihado.