AI / ML Engineer
Posted Updated
Description
Builds the AI-powered features of OpenPHC, the worklist prioritization model, the AI Micro-Coach, and the Data Insights query layer as production features with real statistical rigor behind them, and prepares the technical evidence the Program Director needs for AI governance sign-off.
Responsibilities:
- Worklist prioritization designs and builds the ranking model that determines which patients a frontline worker should follow up with first.
- AI Micro-Coach builds the content model and coaching logic for frontline worker training; the screens themselves are the Senior Full-Stack Engineer's.
- Data Insights builds the natural-language query layer for administrators; the dashboard that consumes it is the Senior Full-Stack Engineer's.
- PII Masking and AI Safety: Ensures no raw PII/PHI is passed to LLM providers or external prompt logs during Micro-Coach or Data Insights processing.
- Responsible AI evidence prepares safety, appropriateness, and failure-mode documentation for each feature, feeding OpenPHC's AI governance sign-off gate.
- Test discipline builds the eval suite for each AI feature and writes automated tests against its documented requirements.
Decision authority:
- Decides model architecture and implementation choices within an approved feature scope.
- No AI feature ships without OpenPHC's sign-off on safety/privacy/cost/data-governance; this gate is prepared for by this role, never exercised by it.
Requirements:
- 5+ years building applied AI/ML products, with at least one shipped to real, non-technical end users.
- Hands-on production experience with LLMs, prompt design, evaluation, and handling failure modes.
- Strong Python and a genuine grounding in statistics/econometrics: this role designs a ranking/prioritization model, not just an API wrapper.
- Able to scope ambiguous, not-yet-well-defined problems into something buildable.
- AI-first, 10x mindset uses AI-assisted coding and eval tooling as the default way of building and iterating on models, expected to ship production AI features at several times the pace of a traditional non-AI-assisted ML engineer without cutting corners on responsible-AI evidence.
Nice-to-have:
- Ranking/recommendation systems experience.
- Any Indic-language NLP or speech work.
- Adaptive learning/micro-learning product experience.
Skills
Sign in to read and join the discussion.