Hyperform Fitness

Hyperform Fitness is a computer vision startup I co-founded that's reshaping how athletes and gym members get coached. The platform uses multi-camera 3D pose estimation to track human movement in real time at 30fps, delivering sub-100ms corrective feedback to lifters during their actual sets — not minutes later in a video review, but right then, in the moment, while the bar is still on their back.

The architecture is edge-first by design. Every frame of inference runs locally on dedicated hardware at the gym, with zero cloud dependency. This matters for two reasons: latency (you can't coach a squat from a server in another state) and privacy (no one wants their body video uploaded to someone else's data center). The system has analyzed over 1 million reps across pilot installations in commercial gyms, rehabilitation centers, and sports medicine programs.

As Co-Founder and Engineering Lead, I architect the inference pipeline, manage a team of interns across software, hardware integration, and client-facing deployment, and actively close deals with fitness facilities and school athletic programs.

The next frontier we're exploring is detecting changes in sleep patterns and vitamin processing through subtle shifts in movement quality — the idea that how you move on a given day reveals far more about your physiological state than wearables can capture.