Member of Technical Staff, Compiler
Summary
Builds a production compiler that turns clinical standards into executable, traceable decision logic for a healthtech team — parsing guidelines, designing logic representations, benchmarking and fine-tuning LLMs, and building deterministic validators with static analysis and formal methods. Core stack: Python, PyTorch, Transformers.
About the Role
Join a healthtech team building a production compiler that turns clinical standards into executable, traceable decision logic. This research-focused role combines experimentation with frontier models and production engineering to improve the reliability and inspectability of clinical decision systems.
What You'll Do
Build pipelines that parse clinical standards and transform them into validated executable modules.
Design and refine a representation for guideline logic, including exclusions, precedence, and recommendations.
Evaluate open and closed models, fine-tuning approaches, and constrained decoding techniques.
Develop benchmarks, datasets, metrics, and adversarial cases to assess clinical fidelity and logical completeness.
Analyze model errors, turn findings into experiments, and ship successful improvements into production.
Build deterministic validators using static analysis, constraint solving, or formal methods to surface ambiguity and contradiction.
Collaborate with clinicians and technical teammates to validate outputs and improve the compiler.
What We're Looking For
At least 3 years of research or industry experience, or final-year PhD or postdoctoral experience, with production engineering work beyond research roles.
Experience shipping ML or NLP systems to production and validating model outputs rather than assuming they are correct.
A bachelor's degree or higher in computer science, machine learning, mathematics, or a related technical field.
Production-level Python, PyTorch, and Transformers experience, plus experience with LLM evaluation, training, or prompt engineering.
Clear technical writing about failed experiments, uncertainty, and results.
Experience in compilers, program synthesis, formal methods, static analysis, IR design, constraint solving, SAT/SMT, or DSLs is valuable.
Healthcare, safety-critical, or regulated-domain experience is valuable, as is familiarity with FHIR or HL7 v2.
Comfort working across disciplines and with ambiguity, and willingness to attend in-person offsites multiple times a year.
Compensation & Benefits
Base salary: $250,000 to $400,000 USD annually. Visa sponsorship is not available.
Location
Remote, based in Canada. Remote candidates in the United States may also be considered.
Skills
As published by ashby · 2 questions
Basics
Name, Email, Resume
Short answers (1)
- LinkedIn Profile URL optional
Pick from a list (1)
- Do you have work authorization for the country this role is based in? optional
