Computer Vision Developer — Real-Time Exercise Tracking (React Native app)

We’re building a fitness app that uses computer vision to count exercise repetitions and track form in real time. We’re looking for an experienced Computer Vision developer to own this feature, starting with push-ups. Accuracy is the #1 priority — rep counting must be precise and reliable. We already have a push-up proof of concept. Your first task is to review it and advise whether to improve it or rebuild from scratch, then deliver accurate, dependable push-up tracking. This first task is time-critical. After push-ups, we’ll expand to more movements — walking lunges, squats, and many others (potentially 10–20+ over time). For the right person, there’s also a large backlog of other app work beyond computer vision. Requirements • Proven, hands-on experience with real-time pose estimation / skeleton tracking (MediaPipe, MoveNet, or similar) • Strong grasp of landmark tracking, rep-counting logic, and how to ensure accuracy and smooth on-device performance • Experience shipping CV features in mobile apps; React Native a strong plus Timeline: Must be able to start next week at the latest. Push-ups are time-critical; no fixed deadline for the additional exercises. Please share examples of similar work and a brief note on how you’d approach this.