Technical Lead AI/ML(Medical Device)

Technical Lead AI/ML Scientist (Hands-on Leadership Role)

San Francisco Bay Area, CA (Hybrid — 3 days on-site)

About the Opportunity

Our client is a well-capitalized, growth-stage medical device company tackling a minimally invasive procedural category with a significant unmet clinical need. Their platform is deployed in the operating room, where AI-driven guidance directly informs surgeon decision-making in real time. The AI/ML program is early and intentional — this hire will shape its technical foundation.

This is the senior technical position within the AI group. The Principal AI/ML Scientist will own the science across the full portfolio: model architecture, dataset strategy, evaluation methodology, on-device deployment, and post-market monitoring. The feedback loop between model work and clinical outcomes is short, and the impact of this role will be visible in the field.

What You’ll Do

Technical Direction & Roadmap

• Set the technical direction for the AI/ML portfolio: model architectures, dataset strategy, evaluation methodology, and on-device deployment patterns.
• Partner with Senior R&D Leadership, Product, and clinical advisors to shape the AI roadmap and make build/buy/partner calls.

Product & Science Leadership

• Lead new product investigations from clinical question through dataset design, baseline model, and deployment plan.
• Translate ambiguous clinical needs into well-scoped ML problems with tangible, measurable success criteria

Team & Mentorship

• Mentor AI/ML engineers at early-career and mid-levels; set the bar for technical work product including dataset standards, reproducible experiments, and defensible model cards.
• Review team output for scientific rigor and reproducibility

Regulatory Ownership

• Author and review the model-validation, generalization, and bias-and-fairness sections of 510(k) and Predetermined Change Control Plan (PCCP) submissions. Partner closely with regulatory and quality teammates to ensure the science holds up under FDA review.

Post-Market Science

• Lead post-market performance investigations once products launch: diagnose drift, propose interventions, and shepherd PCCP-modification cycles.
• Own the long-tail scientific accountability that extends well beyond initial clearance

External Representation

• Represent the AI program at clinical conferences, academic medical center collaborations, scientific advisory interactions, and technical conversations with FDA as needed.

Required Qualificaitons

• PhD, or MS plus equivalent industry experience, in Computer Science, Electrical Engineering, Biomedical Engineering, Applied Mathematics, or a related technical field.
• 8+ years of applied machine learning or computer vision experience, with at least one shipped medical device or safety-critical AI product where you led the scientific work.
• Deep working knowledge of modern computer vision — detection, segmentation, tracking, and geometric reasoning — and the training and evaluation workflows that go with them.
• Direct experience leading the AI/ML scientific content of FDA submissions (510(k) AI/ML, PCCP) or equivalent regulatory work in another regulated domain.
• Strong Python and deep familiarity with PyTorch (or TensorFlow); comfort with experiment tracking, dataset versioning, real-time inference, and model monitoring.
• Demonstrated technical mentorship: prior experience setting the technical bar for a small team and reviewing their work for scientific rigor.
• Excellent written communication for regulatory, clinical, and quality

Preferred Qualifications

• Background in endoscopy, urology, or other intra-luminal surgical AI.
• SLAM, structure-from-motion, 3D reconstruction, or geometric computer vision applied to surgical or endoscopic video.
• Real-time inference optimization for medical-grade embedded hardware: TensorRT, ONNX, quantization, knowledge distillation.Authorship on Class II AI/ML 510(k) clearances or comparable submissions.
• Active publication record in venues such as npj Digital Medicine, MICCAI, MIDL, IEEE TMI, J Endourol, BJU Int, or similar.
• Familiarity with IEC 62304, ISO 14971, and AA MI TIR34971.