Computer Vision Engineer

About MacV AI At MacV AI, we build real-time Vision AI systems for industrial environments — processing live video streams to understand what's happening on the ground and prevent costly or dangerous outcomes. We're a small, passionate team working together to leverage cutting-edge technology to help build safer workplaces. About the role We're looking for a Computer Vision Engineer (2–4 years experience) who has hands-on experience building and deploying production-grade vision systems. You will work directly with the founders on core product systems across two main tracks: Industrial Safety Systems: Real-time object detection for safety use cases, low-latency inference, and alert generation on live video streams Retail Intelligence: People detection, multi-object tracking, cross-camera re-identification, and demographic estimation What you'll do Design and deploy end-to-end computer vision systems for real-world environments Train, fine-tune, and optimize models for detection, tracking, re-identification, and feature extraction Build and optimize real-time inference pipelines (latency, throughput, GPU utilization) Work on multi-camera systems including synchronization and identity tracking across feeds Integrate models into backend systems (APIs, databases, alerting pipelines) Evaluate trade-offs between accuracy, speed, and cost in production environments Experiment with state-of-the-art approaches (e.g. VLMs) where relevant to real-time systems Requirements 2–4 years of experience in a computer vision / ML engineering role Proven experience building/deploying CV pipelines in production (not just academic or side projects) Strong proficiency in Python and PyTorch (or TensorFlow) Hands-on experience with object detection, multi-object tracking, model optimization Experience building or contributing to real-time video pipelines Familiarity with backend systems (FastAPI/Django) and databases (PostgreSQL/MongoDB) Strong debugging skills in real-world scenarios (data issues, latency, deployment failures) Strong Plus Experience with NVIDIA DeepStream, TensorRT, ONNX Runtime Work on multi-camera tracking / re-identification systems Experience deploying on edge devices (Jetson, GPU servers) What we offer High ownership - you'll work on systems that go live with real customers Meaningful ESOPs (employee stock options) designed to reward early team members as the company scales Direct exposure to real-world deployment challenges Opportunity to grow into a senior / lead role quickly Tech Stack Python, PyTorch, FastAPI, PostgreSQL, MongoDB, YOLO, VLMs, NVIDIA DeepStream, AWS / Azure