Technical Product Manager, Trust & Safety
Core
Define and ship data products, RL environments, and evaluation frameworks for frontier AI labs to ensure safe training and deployment of agentic models.
Role type
Senior Technical Product Manager (AI Safety & Trust)
Builds
Data products for model training/evaluation, sandboxed RL environments with adversarial traps, safety rubrics, and metrics harnesses.
Domain
Generative AI, AI Safety, Trust & Safety, Agentic AI
Deliverable
production ML models
Required skills
Product management for ML/AI, RLHF/RLAIF design, agentic AI architecture, AI safety concepts (reward hacking, deceptive alignment), adversarial thinking, technical spec writing, metric definition.
Preferred skills
Python, Docker/VMs, browser automation, red-teaming experience.
Technologies
Python, Docker, VMs, browser automation tools
Responsibilities
Translate lab research goals into product specs for safety training and RL environments; design sandboxed environments with misalignment and cyber traps; define safety rubrics and reward signals for agentic tasks; own evaluation metrics and track adversarial success rates; prioritize roadmap items based on threat landscape analysis; partner with researchers, engineers, and delivery teams.
Seniority
Senior, hands-on IC
