Senior Staff Software Engineer - Health Intelligence & Decision Engine

Prime Health Technologies is building ELIA, a personalized health intelligence system that turns trusted science and personal health data into practical actions. The Senior Staff Software Engineer will build its Health Intelligence & Decision Engine, translating expert-authored guidance into reliable, testable production software and personalized, explainable recommendations.

Responsibilities:

• Write, debug, refactor, test, and ship production TypeScript code for ELIA's Health Intelligence & Decision Engine. Use AI coding agents as accelerators, while independently validating and owning the resulting software
• Translate scientific documents, tables, and formulas into clear specifications, tools, and skills. Use AI-assisted analysis and human review to surface gaps, conflicts, and ambiguity; resolve scientific questions with their owners before implementation
• Implement decision logic that determines whether and when to engage a user and selects the most appropriate recommendation from available information. Make precedence, safety exclusions, preferences, data confidence, freshness, and missing-data behavior explicit
• Combine deterministic calculations with LLM interpretation and generation. Validate model-derived inputs and filter proposed options against contraindications and user preferences. Define evals, safe fallback or no-recommendation behavior, and clear limits and retirement conditions for temporary fallbacks
• Maintain traceability from versioned scientific sources and approved specifications through implemented rules, tests, and operational decisions. Capture the inputs and model outputs actually used, rule and formula versions, software release, and decision reasons in event-sourced records. Support historical decision replay and comparison with revised rules while preserving the original record
• Direct multiple AI agents through specification, implementation, code review, testing, and integration in rapid, small batches, parallelizing where practical. Pair AI code review with human judgment. After each build round, assess requirements coverage, test coverage and results, errors, gaps, conflicts, bloat, and architectural drift. Use the findings to improve specifications, agent instructions, task breakdown, and review checks, tracking delivery quality and rework over time
• Bring independent engineering judgment to design reviews with the Chief AI & Systems Architect. Challenge assumptions constructively, propose alternatives, make tradeoffs clear, and translate decisions into shipped code
• Build scenario, regression, safety, and interface tests. Agree tested contracts with the Interaction + Harness team, learn adjacent components, and share designs and reviews to keep engine and interaction work aligned
• Align with the Chief AI & Systems Architect on the engine's detailed design, initial priorities, and integration contracts; turn a representative source set into reviewed specifications with an explicit list of unresolved questions
• Personally ship one useful end-to-end capability through ELIA, with production code, safety checks, scenario tests, evals for any LLM behavior, and a demonstrated trace from a user-facing decision back to its inputs, implemented rules, approved specifications, and scientific sources
• Demonstrate a repeatable agent-assisted delivery process, decision replay and versioning, and a practical plan for the next capabilities and temporary fallbacks

Requirements:

• 7+ years building software, including systems you designed from the start; strong TypeScript and the ability to explain architectural boundaries and tradeoffs
• Substantial experience with rules, policy, or decision engines that turn expert-authored guidance into tested, version-controlled rules and reproducible decisions
• Product judgment and experience with specification-driven development: translating goals into requirements, interfaces, acceptance criteria, and incremental releases. We value the principles across different approaches; no particular framework is required
• Experience leading delivery through a software team or multiple AI agents, with hands-on proficiency directing coding agents and critically reviewing their output
• Be prepared to demonstrate your agentic development process from specification through review, testing, and release, including how you evaluate and improve successive build rounds
• Production experience with LLM APIs, tool use, and agent workflows, with practical proficiency designing and interpreting evals
• Strong testing and debugging skills, including safety constraints, failure cases, and tracing problems across systems
• High degree of proficiency in written and spoken English, sufficient to accurately interpret complex scientific material, resolve ambiguities with its authors, and write precise implementation specifications
• Candidates must demonstrate their understanding of the source material and explain their decisions during practical interview exercises
• Examples you can discuss of architecture you designed and specifications you carried through implementation, including key tradeoffs and validation evidence
• Redacted or personal work is welcome
• Healthcare experience is strongly preferred
• Other useful experience includes clinical decision support, regulated or privacy-sensitive systems, event sourcing and audit trails, customer-controlled infrastructure, and early-stage product development