Location: Penang (hybrid) Team: Sports-AI (badminton performance, broadcast analytics, and 3D R&D) Mission We turn video (and sensors) into real-time coaching, broadcast-ready analytics, and a millimeter-accurate 3D "digital twin" of on-court action. We're hiring one posting for multiple openings, you'll own a track based on your strengths: Choose your track (you'll focus on one) AI Coach (CV/ML ) - 2D/3D human pose estimation, fine-grained action recognition (smash, drop, clear, net), robust player & shuttlecock tracking, convert joint angles/shot placement/coverage into simple coaching advice and clear visualizations. Broadcast Analytics & Viz - Low-latency tracking from broadcast feeds; real-time metrics (shuttle speed, player speed, distance, rally stats); heatmaps, shot charts, rally trajectories; automated event detection (smashes, long rallies) and stunning on-screen overlays. 3D Multi-Camera R&D (Senior) - Multi-camera calibration & time sync, SfM/MVS 3D reconstruction for players & shuttle path, C++/CUDA real-time 3D tracking, advanced 3D pose/kinematics (angles, angular velocity, forces), and architecture of the full multi-cam stack. What you'll do (core) Design, train, and optimize CV/DL models for real-time performance. Build reliable video pipelines (on-device or cloud) and profiling to hit latency targets. Collaborate with coaches, broadcast, and product to ship features that athletes & fans love. Must-have (vary by level) Mid level: 3+ years in CV/ML; strong Python; PyTorch or TensorFlow; OpenCV; hands-on with pose estimation (MediaPipe/OpenPose), object detection/tracking (YOLOv8, DeepSORT, TrackNet) or action recognition. Senior 3D track: 5+ years in 3D geometry; expert C++ & CUDA (mandatory); deep grasp of camera models, epipolar geometry, triangulation, SfM/MVS; proven GPU real-time optimization. Nice to have TensorRT/CoreML, FFmpeg/GStreamer, WebRTC/RTSP, AWS/GCP/Azure, multi-camera rigs & time-sync (PTP), motion capture (Vicon/OptiTrack), NeRF/differentiable rendering, sports analytics/badminton domain knowledge, front-end overlay collab. How to apply Apply here on LinkedIn and send your CV + links (GitHub/portfolio/Papers) to [[HIDDEN TEXT]] with subject: "Sports AI - CV/3D Engineer (Track: Coach / Broadcast / 3D)". Tell us your track, notable projects, and earliest start date. Keywords (ATS): Python, C++, PyTorch, TensorFlow, OpenCV, FFmpeg, CUDA, TensorRT, YOLOv8, DeepSORT, TrackNet, MediaPipe, OpenPose, SfM, MVS, epipolar geometry, triangulation, kinematics, WebRTC, RTSP, AWS, GCP, Azure. Show more Show less