The company is a healthtech team building clinician-facing decision support for Canadian public healthcare.
You will create reinforcement learning environments and evaluations that make clinical AI systems more dependable, working alongside the researchers and clinicians who check the outputs.
Responsibilities
Create reinforcement learning environments for demanding clinical reasoning, covering multi-step work, tool use, state changes, and measurable criteria for success.
Shape and tune reward functions, grading rubrics, and LLM judges, then probe them for gaming by the model and for scores that mislead.
Take ownership of task design, data generation, environment build, evaluation, and handoff to the teams that train models.
Run post-training and agent evaluation experiments, and use the failures to refine the environments.
Work with clinicians and research leads to confirm model outputs and the quality of evaluations.
Requirements
Either a PhD in computer science, machine learning, or a related field, or 3 years of full-time reinforcement learning research or post-training experience.
At least 3 years of research or industry experience, or currently a final-year PhD student or postdoctoral researcher.
A bachelor's degree or higher in computer science, machine learning, mathematics, or a related technical field.
Hands-on work building reinforcement learning environments, post-training systems, or agent evaluations, plus production experience with Python, PyTorch, and Transformer models.
Experience owning reward functions, grading rubrics, or LLM judge calibration, and writing up failed experiments and open uncertainty clearly.
Deep research background, experience at a frontier AI lab or leading AI company, comfort with ambiguity and cross-disciplinary work, and careful checking of model outputs that look plausible.
Useful: clinical AI experience, knowledge of information retrieval design, constraint solving, SAT/SMT, or domain-specific languages.
Conditions
Base salary of $250,000 to $400,000 per year.
Visa sponsorship is not offered. You must already be authorized to work in Canada or the United States.
Remote role, with Montreal, Quebec, Canada as the main location. Candidates working remotely in Canada or the United States can apply.