Senior Ai Engineer Computer Vision

## About Clutch Clutch is on a mission to build a global community for racket sports players - powered by an AI camera infrastructure that helps players re-live their best moments on court and track their progress with game analysis. Our initial focus is padel, the fastest growing sport in the world. We've spent the past year designing and building our own camera, the Clutch Cam, and we've proven the product works in the field - with presence in over 100 clubs across the US, Europe, and the Middle East. Now we need to make the AI bulletproof at scale. We're a small, ambitious, and fully remote team - and we plan to keep it that way for the foreseeable future. ## About the role We're looking for a technical leader to own and evolve our AI pipeline and join us on our mission to build a global racket sports community. As Clutch's Senior AI Engineer, you'll be responsible for performance and accuracy of our main backbone - AI pipeline. Our current pipeline runs on GCP and processes video in real time from the Clutch Cam, our custom-built camera hardware. You'll be working on some genuinely hard problems: player tracking and re-identification, shot classification, and ball tracking - the ingredients behind both automated video production and game analysis. Your work would touch upon all parts of ML lifecycle. That means building and maintaining real-time video understanding pipelines, optimizing models for low-latency inference on cloud infrastructure, and managing everything from dataset design and annotation workflows to model versioning, model cards, and production monitoring. This is a hands-on role. We're looking for someone who thrives in that environment: someone who can go deep on a hard CV problem in the morning and think about pipeline architecture in the afternoon. And if you play padel, tennis, or any racket sport yourself - you'll feel the impact of your work every time you step on court. In your first 6 months, you'll focus on accuracy improvements of our most crucial models in event detection and ball trajectory subsystems. ## Hiring process A few conversations and a technical challenge. We keep it straightforward and respectful of your time. ## Minimum qualifications - 3+ years of industry experience applying Computer Vision to real-world problems - Experience working with at least one of the following: action recognition, temporal action localization, and object tracking - Experience with video understanding pipelines - multi-frame models, real-time stream processing, or similar - Experience with technologies like PyTorch, TensorFlow, Docker and OpenCV - Experience deploying AI pipelines in the cloud, eg. AWS, Azure or GCP - Strong Python skills and a track record of writing robust, production-grade code ## Bonus qualifications - Experience with model optimization for real-time inference on cloud machines - quantization, ONNX - Experience monitoring models in production, including tracking accuracy drift and setting up performance alerting - Experience with MLOps practices including dataset versioning, model versioning and testing - Experience leading annotation workflows and dataset design, including corner case identification and systematic model evaluation - Experience with model deployment on edge devices - A passion for racket sports