Computer Vision / Machine Learning Engineer

Computer Vision / Machine Learning Engineer (Contract) Contract Length: 6 months Work Arrangement: Hybrid (onsite 3 days/week in Greater Boston) Computer Vision / Machine Learning Engineer to support an active, production-focused initiative. This role is centered on improving detection and tracking accuracy for object-counting systems operating in real-world, high-throughput environments. The ideal candidate has hands-on experience building and deploying computer vision models, working with multimodal sensor data, and optimizing inference pipelines for edge deployment. What You’ll Work On • Enhancing object detection, segmentation, and tracking accuracy in operational systems • Developing and deploying models into validation and pre-production environments • Improving real-time performance and reliability on edge hardware Key Responsibilities • Model Development: Train, validate, and deploy computer vision models, with an emphasis on instance segmentation • Tracking & Fusion: Implement tracking approaches that combine color and depth data to maintain object persistence across frames • Data Quality: Support data curation efforts and audit external annotations to ensure high-quality ground truth • Performance Optimization: Tune inference pipelines for low-latency execution on edge platforms Technical Environment • Computer Vision & ML: Instance Segmentation, Object Tracking • Sensor Data: RGB + Depth (basic multi-sensor fusion) • Edge & Optimization: NVIDIA-based edge hardware, TensorRT or similar acceleration tools Qualifications • Proven experience delivering production-grade ML or CV systems • Strong software engineering fundamentals (version control, testing, CI/CD) • Experience deploying models beyond experimentation into real environments • Ability to meet strict accuracy and performance benchmarks • Comfortable working within cloud-only data environments with controlled access policies Nice to Have • Experience optimizing models for edge or embedded systems • Familiarity with real-time or near–real-time vision pipelines • Background in industrial, robotics, or high-volume operational settings