YEAR:

2026

2026

EXPERIENCE:

Swift, Python, Core ML/AI, PyTorch, iOS

Swift, Python, Core ML/AI, PyTorch, iOS

PaddleIQ logo.

AI Table Tennis Coach

PaddleIQ is my iPhone app for table-tennis match analysis, now in private beta. I built the SwiftUI app and converted the existing TrackNet model from Python for Core ML. The app also uses a Kalman filter in its tracking pipeline.

Tracking pipeline

TrackNet processes video frames, and a Kalman filter smooths the tracking output. The app analyzes iPhone video and ball-contact audio to produce match stats, highlights, replay clips, and coaching checkpoints. The beta supports recording or importing a table-tennis match.


PaddleIQ matches screen


PaddleIQ match recording setup


PaddleIQ 3D stroke replay


PaddleIQ stroke coaching checkpoints

Results and limits

PaddleIQ is in private iPhone beta. Players can record or import a match, review match and stroke stats, watch replay clips, and choose coaching checkpoints for practice. The product site shows six app screens. On-device speed and tracking accuracy have not been benchmarked here.