AI Deployment Manager

Responsibilities:

• The deployment. Run pilots, take clients live, stabilise them and lead their expansion. You are accountable for the client's outcome end to end, and you are the person who knows where every client stands today.
• The translation. Convert client needs into clear, prioritised asks for our teams, and convert our progress into updates the client can act on. Nothing gets lost between the client and the people building for them.
• The resolution. Trace an issue across data, workflow, model and people, route it to the right owner, and drive it to closure. You report problems solved, not problems found.
• The tooling. Prototype agents, prompts and automations that remove manual steps from this job, deploy them, and measure the impact. Where a process is repeated, you are expected to build rather than repeat.
• The product. Turn patterns you see across clients into product features and process changes, so that every client after yours runs on what you learned.

Requirements:

• 2+ years of experience, with shipped AI work. At least one AI-driven solution you built and put in front of real users: an agent, an automation, an internal tool or a product feature. You are hands-on with prompting, agents and evals, and you can walk us through what you built, how you tested it, and what broke.
• A fast learner. Healthcare billing and medical coding are deep domains. You do not need to know them today, but you need to pick them up quickly and enjoy doing so. We are betting on learning speed over prior domain exposure.
• A structured thinker. You break messy problems down MECE (mutually exclusive, collectively exhaustive), so nothing is missed, nothing is counted twice, and you can hold that structure while the details keep changing.
• Comfort with data. SQL, Python or advanced spreadsheets, enough to answer your own questions without waiting on someone else.
• Clear communication, written and spoken. You will talk to clients, engineers, coders and founders in the same day, and each of them needs a different version of the same truth.
• Comfort with context switching. This role moves between a client call, a stuck batch and a prompt that is not working. It is not a role with long quiet blocks, and the people who thrive in it find that energising rather than draining.
• High ownership and a problem-solving nature. You start with "how do I fix this", and you want the zero-to-one journey with the mess that comes with it.
• Exposure to healthcare, billing or US insurance. A prior client-facing or operations role at a startup. Familiarity with agent frameworks, structured outputs and tool calling, or LLM evaluation tools.