Lead Computer Vision/ML Engineer

About the role

Norbert-health is seeking a lead deep‑learning engineer to drive the development of its autonomous healthcare‑robot sensing platform. You will own the computer‑vision stack from research through production, enabling real‑time vital‑sign extraction on edge devices.

Key responsibilities

• Design, fine‑tune, and deploy computer‑vision models (YOLO, InsightFace, MediaPipe, facial landmark detection, object tracking, pose estimation) for real‑time inference on the edge.
• Optimize models for embedded deployment using quantization, pruning, TensorRT, and NVIDIA Triton.
• Build and maintain MLOps pipelines for training, validation, and performance monitoring.
• Develop video‑processing pipelines that combine classical signal processing with ML‑based vital‑sign extraction.
• Establish engineering best practices, reduce technical debt, and contribute to the overall architecture of the vision stack.

Required profile

• Master's or PhD in Machine Learning or Computer Vision.
• Strong fundamentals in data structures, CV algorithms, and systems programming.
• 5+ years experience deploying computer‑vision models to production on resource‑constrained devices.
• Ability to work independently, solve complex problems, and drive projects to completion.

Required skills

• Expert C++ programming (critical for edge deployment).
• Proficient Python for ML experimentation and tooling.
• Experience with PyTorch and model optimization for edge AI.
• Knowledge of NVIDIA Jetson platform, TensorRT, and Triton Inference Server.
• MLOps tools for experiment tracking, model versioning, and performance monitoring.

What we offer

• Real impact: your code directly supports patient care today.
• High autonomy and technical ownership over the computer‑vision architecture.
• Work at the intersection of cutting‑edge AI, edge computing, and healthcare.
• Talented, diverse, international team with offices in Brooklyn and Paris.
• Competitive salary and equity in a mission‑driven, transparent culture.