AI Engineer (Healthcare) | San Francisco | $220k + equity

I'm working with a rapidly growing AI company in San Francisco that is building its product around autonomous AI agents. They're looking for an AI Engineer to own the LLM-powered systems at the heart of the product — from agentic workflows and evaluation infrastructure through to retrieval, knowledge systems, and AI features used by customers and internal teams. This is an early-stage, high-ownership role. You'll work directly with the founders, have significant influence over the technical direction, and help define how the company builds reliable AI systems as the technology evolves. Responsibilities: Build agents that do real work Design and ship LLM-powered systems that can take complex, multi-step tasks from start to finish — deciding what needs to happen, taking actions across internal and external systems, evaluating their own work, and knowing when to involve a human. Own AI quality and evaluation Build the infrastructure needed to understand how well the AI is actually performing. You'll develop meaningful evaluation frameworks, monitor production behaviour, identify regressions, and use real-world signal to continuously improve prompts, retrieval and models. Build the knowledge layer Design the retrieval and knowledge systems that agents rely on to reason accurately. A major part of the challenge is ensuring that information stays reliable, current and correctly versioned as underlying sources change. Turn AI into product Build copilots and AI-powered features that make users meaningfully more productive. You'll need strong product judgment around when the system should act autonomously versus when it should ask for human input. Ship, measure, iterate You'll be close to production. Deploy systems, observe how they behave with real users, and iterate quickly based on what actually happens rather than relying solely on offline benchmarks. Shape the roadmap Work directly with the founders to decide what to build next. You'll bring a practical understanding of what today's models can and can't do, and help identify where the technology can create genuine product leverage. Build for scale Make the architectural decisions needed to support significant growth while keeping systems reliable, maintainable and easy for future engineers to build on. Quaifications: You've built and shipped LLM-powered systems or agents into production , rather than working primarily on demos or prototypes. You're genuinely fluent in modern LLM systems and can reason practically about prompting, tool use, context management, retrieval, latency, cost and failure modes . You've thought seriously about evaluation — defining metrics, collecting production feedback and using that signal to systematically improve AI quality. You've operated AI systems in production and understand the challenges of monitoring, debugging and maintaining reliability when real users depend on them. You're a strong software engineer who writes maintainable code and takes ownership of what you build. You're comfortable operating in an environment where the playbook hasn't been written yet and enjoy figuring out what should be built rather than simply executing a predefined roadmap. If this sounds like something of interest, feel free to apply below or reach out to me at reece@harrisonclarke.com Show more Show less