Senior Computer Vision Engineer

We’re partnering with a fast-scaling startup in SF that is fundamentally changing how physical materials are tracked and measured using real-time 3D computer vision. If you're passionate about bringing SOTA 3D vision algorithms out of the research lab and into high-impact, real-world industrial environments, this role gives you full ownership over the core perception engine. They are looking for an experienced Senior Computer Vision / Perception ML Engineer to own the perception engine that converts raw stereo camera captures into accurate, production-grade 3D measurements for industrial applications. This role sits directly at the core of our tech stack. You will be responsible for taking complex research pipelines—covering stereo reconstruction, point-cloud cleanup, depth estimation, and 3D volumetric metrics—and turning them into robust, production-ready algorithms. What You Will Own The core end-to-end CV/ML pipeline: stereo reconstruction, point-cloud processing, segmentation, ground-plane estimation, stitching, and volume calculation. System calibration, confidence metrics, failure-mode analysis, and 3D reconstruction quality. Bridging the gap between prototype research and deployable, real-world sensor performance. What We’re Looking For Experience: 4+ years of hands-on experience in computer vision, robotics perception, or ML engineering. 3D Vision Focus: Deep expertise in 3D geometry, camera calibration, stereo/depth estimation, point clouds, and segmentation. Core Tech Stack: Proficient in Python, PyTorch, OpenCV, and Open3D. Real-World Engineering: Proven track record of debugging noisy sensor data and shipping CV systems in real-world environments. Education: MS or PhD in Computer Vision, Robotics, or a related field preferred. Nice-to-Haves Hands-on experience with CUDA, SLAM, photogrammetry, or geospatial/volumetric measurement. Experience with cloud GPU infrastructure (e.g., Modal). Prior perception experience in robotics, autonomous vehicles, or industrial inspection. #J-18808-Ljbffr