SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Translucent is building an agentic AI platform designed exclusively for healthcare finance. The company was founded in 2024 and is backed by GV, NEA, FPV, and Virtue. They've already deployed their product with healthcare organizations managing over $5 billion in combined revenue.
As an AI Engineer, you'll work hands-on across R&D and platform engineering to improve the underlying agent capabilities that every team builds on. Rather than configuring agents for individual customers, you'll focus on standardizing tool surfaces, making the agent platform integration-ready, and owning the evaluations and context engineering that make the system reliable for healthcare finance.
Key responsibilities include:
- Standardizing tool and skill surfaces by defining contracts (MCP-style interfaces) that allow continuous capability additions without destabilizing the platform
- Building connective tissue so agents, tools, context, and data compose cleanly across products
- Owning production-grade evaluation harnesses, replay systems, and benchmarks for agentic AI; accuracy is critical in healthcare finance
- Establishing best practices for grounding agents in customer business rules and data, and maintaining control loops for reliability
- Fine-tuning in-house and open-source models, benchmarking against frontier baselines, and making build-vs-buy decisions
- Turning core capabilities into reusable components that translate across features and products
You'll need 3+ years of software engineering experience with meaningful production Python work, direct experience shipping LLM-powered or agentic systems in production (not prototypes), and hands-on experience with modern agent frameworks like LangGraph, Google ADK, LlamaIndex, or Claude Agent SDK. You should have built production evaluation harnesses and benchmarks that gate real releases, experience grounding agents through retrieval and context layers, and comfort with SQL, BigQuery, embeddings, and vector search. A bias toward shipping working V1s quickly and iterating is essential.
Deep expertise in at least one (ideally two) of these areas is preferred: tool-surface and connector platform engineering, harness and loop engineering, open-source and in-house model fine-tuning, or production context-layer engineering. Experience with healthcare finance, accounting, or structured financial data is a nice-to-have.
