Senior AI Engineer

Why Translucent

Healthcare providers operate on razor-thin 2-5% margins despite $2.5 trillion in annual expenditures. Finance teams struggle with spreadsheets, manual data pulls, and disconnected systems. Translucent builds agentic AI platforms exclusively for healthcare finance, deploying AI Agents operating 24/7 across specific data, business logic, and workflows. Founded in 2024 with backing from GV, NEA, FPV, and Virtue, the company already serves healthcare organizations managing over $5 billion in combined revenue.

About the Role

Build healthcare's AI financial analyst. Agents autonomously plan multi-step workflows, select and invoke tools, retrieve context from heterogeneous data sources, execute SQL, generate visualizations, and synthesize findings. Own the full agentic feature lifecycle, from tool design through production monitoring.

What You'll Do

• Design agentic workflow graphs in LangGraph with nodes, conditional routing, state management, and error recovery
• Build composable agent tools (SQL executor, chart generator, file parser) with typed schemas
• Design retrieval pipelines: semantic search, hybrid retrieval, re-ranking, context-window optimization
• Own semantic layer mapping natural language to BigQuery schemas and business rules
• Author and evaluate prompts for Gemini models with structured outputs and tool-calling specs
• Build evaluation frameworks with unit tests, LLM-as-judge scoring, and regression suites
• Instrument agent runs with tracing and observability

What You Bring - Must Have

• Shipped agentic architectures: state machines, tool routing, planning loops
• Tool-calling pipelines with schemas, selection, invocation, validation, and chaining
• RAG pipelines, vector + keyword hybrid search, re-ranking, embedding strategies
• Prompt design, structured output extraction, guardrails, systematic evaluation
• Production-grade Python 3.10+ with async/await and Pydantic (~80% of work)
• Analytical SQL fluency; BigQuery experience preferred
• pytest, mocking LLM responses, evaluation-driven development

Bonus Skills

• GCP services (BigQuery, Vertex AI, Cloud Storage, Cloud Run)
• Text-to-SQL or semantic layer systems
• MCP (Model Context Protocol)
• Vector databases (Pinecone, Weaviate, pgvector)
• Healthcare data or enterprise multi-tenant SaaS

Mindset

Comfortable with ambiguity in experimental agentic AI. Owner mentality from design through production. Architecture thinking for refactoring into composable tools. Customer-first approach solving real healthcare problems.

Compensation and Benefits

Competitive salary and equity at early-stage company. High-impact role shaping core agentic AI platform. Hybrid NYC with flexibility. Salary $200,000 - $250,000.