About the Role
You will build post-training and agent systems within strict data, deployment, access, auditability, and human-authority boundaries. You will create private improvement loops, permissioned and traceable agents, and evaluations for reliability, calibration, recovery, and compliance.
Requirements
•Experience in reinforcement learning, post-training, private ML systems, high-assurance evaluation, agent safety, regulated-domain ML, public-sector technical work, or constrained deployment
•Experience building and debugging model-training, post-training, agent, or evaluation loops
•Experimental judgment, including baselines, trace inspection, and failure attribution
•Python
•Ability to build reproducible artifacts for researchers and engineers
•Ability to own ambiguous problems and communicate with partners
Responsibilities
•Build end-to-end improvement loops inside operator data and infrastructure boundaries
•Develop agents with explicit permissions, evidence trails, bounded tools, and human checkpoints
•Design evaluations for reliability, calibration, recovery, and compliance with real procedures
•Study how privacy, access, and infrastructure constraints affect learning
Benefits
•Flexible work
•Visa sponsorship and relocation support
•Team off-sites and conference attendance
