The role
- COMP
- $250K - $400K
- EQUITY
- Competitive
- LOCATION
- Canada
- WORKPLACE
- Remote
- EXPERIENCE
- 3+ years
- VISA
- None
- STACK
- Python, PyTorch
- INDUSTRY
- AI, Healthcare
/roles — ROLE_322
Member of Technical Staff, Compiler (Remote)
Remote clinical AI team converting medical standards of care into verified, executable decision logic
The company
Small clinical AI research company turning evidence-graded medical guidelines into deterministic, traceable decision tools that clinicians already use in production.
- STAGE
- seed-stage
- FUNDING
- seed funding
- TEAM
- founding team of <10
- FOUNDED
- 2024
JD — the work
About the role
The work is research, but with real production consequences. The company turns clinical care standards into decision logic that is schema-enforced, verified to be exhaustive, and traceable back to the source text that justifies it, and that system is already live in public health care. You would probe the limits of frontier models at pulling out, structuring, and formally proving clinical logic, run the experiments that show which methods hold up, and ship the best one into the pipeline. The group is tiny, works remotely from anywhere in North America, and gathers in person several times a year.
What you'll do
- Build the pipeline that turns clinical guidelines into runnable modules, using layout-aware document models for parsing, then transformation and validation
- Design and refine the target syntax that encodes guideline logic, covering precedence, exclusions, and recommendations
- Compare open and closed models, fine-tuning, and constrained decoding through careful experiments
- Co-design adversarial test cases, datasets, and metrics that gauge clinical fidelity, completeness, and recommendation accuracy
- Write deterministic validators using formal methods, constraint solvers, or static analysis to expose ambiguity and contradictions
- Treat each model failure as a hypothesis to test, feeding results back into system design
- Work hand in hand with the founders and with clinicians whose professional licenses ride on the results
What they're looking for
- 3+ years of experience in applied ML or AI research, with strong Python and PyTorch
- Research background in clinical AI, ideally with publications
- Hands-on LLM experimentation such as fine-tuning, constrained decoding, or benchmark design
- Conviction that clinical decision logic must be inspectable, plus a desire to move past demos
- Comfort working fully remotely with periodic in-person offsites