Job description
Role Overview: We are seeking a highly skilled AI/ML Engineer to lead the development, scaling, and productionization of our advanced AI Agent ecosystem. In this role, you will be responsible for orchestrating multi-agent LLM systems and developing machine learning models to analyze complex clinical data for our maternal-care platform.
You will directly oversee and mature three critical intelligent agents: Agent 1 production billing reconciliation and payer eligibility), Agent 2 navigation automation), and Agent 3 clinical pattern recognition executing against an 11K+ escalation corpus).
Key Responsibilities
Agent Orchestration: Design, build, and optimize multi-agent workflows using TypeScript/Node.js to call enterprise LLM APIs. ML Pattern Recognition: Develop and train specialized Python-based machine learning pipelines to ingest and detect anomalies, trends, and risk indicators within an 11K+ clinical escalation corpus.
Agent Lifecycle Management: Maintain and iteratively improve Agent 1 (cross-reconciliation across PCM/BHI/RPM/CCM), advance Agent 2 through its deployment phases, and mature Agent 3 from build to production readiness. Data Pipeline Integration: Work closely with the data engineering team to process structured and unstructured data via Snowflake (Snowpark / Python APIs) and ensure data compliance with HIPAA standards for handling Protected Health Information (PHI).
System Performance & Evaluation: Establish strict evaluation frameworks (evals) for LLM outputs to guarantee clinical safety, accuracy, and mitigation of hallucinations in triage recommendation queues.
Requirements
Languages: Advanced proficiency in Python (for ML data science workloads) and TypeScript / Node.js (for backend orchestration and API integration). AI/LLM Frameworks: Strong experience with LLM orchestration frameworks (e.g., Lang. Chain, Llama. Index, or Lang. Graph) and commercial/open-source LLM APIs. Machine Learning NLP: Deep understanding of Natural Language Processing (NLP), text embedding generation, vector databases, and pattern-recognition techniques applied to unstructured text datasets.
Data Stack: Hands-on experience with Snowflake and Snowpark using Python APIs. Healthcare Domain (Highly Preferred): Familiarity with US healthcare compliance, HIPAA data privacy requirements, and navigating clinical nomenclature.