AI Engineer

At Rayn, we don’t just work—we innovate, collaborate, and create solutions that leave a lasting impact. As part of our team, you’ll have the opportunity to shape the future of a purpose-driven organization that is redefining how technology can address societal & public health challenges.
This is a builder role. You run each engagement yourself: the first client call, the scoping, the build, the launch and the fixes after launch. Engagements are short and have a clear end. If what you ship is wrong or unreliable, fixing it is your job.

What you will bring to Rayn as an AI Engineer

• Proposal engines that find government and enterprise tenders, check fit, and draft compliant bids.
• AI interviewers that run hundreds of research interviews at once and turn them into clear findings.
• Account research tools that pull data from many sources, enrich it and draft personalised outreach.
• Voice agents that handle real phone calls end to end and connect to a client's CRM and internal systems.
• Speech models and voice apps for local languages, so people who can't read can use digital services by speaking.
• Healthcare AI products that sit inside clinical and admin work, such as intake, prior authorization and patient follow-up.
• Sit with clients, learn how the work actually gets done, and write down what needs building.
• Find the slowest step and measure it before you write any code.
• Decide what to build, what to buy and what to skip, and explain why in one page.
• Ship to production, with tests for correctness, security and cost.
• Build so the next client with the same problem gets it faster, and leave docs good enough for someone else to take over.
• Build our own sales tooling: account research, scoring and outreach agents.

Who We’re Looking For

• 3 to 5 years shipping software that people rely on, including time at a startup or in a client-facing team.
• At least one AI agent in production that other people used for weeks. You can tell us how you tested it, how you monitored it and what it cost to run.
• A clear sense of when a simple rule-based workflow is enough and when you actually need an agent.
• Strong Python or TypeScript, plus APIs, webhooks and SQL.
• Safety and data governance as part of the job: guarding against prompt injection, data leaks and bad output, and handling client data with proper access control and audit trails.
• Careful release habits: tests, gradual rollouts and a way to roll back.
• Strong communication with customer stakeholders, engineers and product teams, in written and spoken English.
• A history of starting things nobody asked you to, and finishing them.

Nice to Have

• You have built evals and guardrails that caught an agent going wrong before a client did, and you can explain how you decided what "good" looks like.
• You understand long-running, multi-step agent tasks: where they drift, loop or lose the thread, and how to keep them on track.
• AI built into real healthcare workflows, at the depth of today's leading AI health products: pulling patient records from EHR systems over FHIR, working with HL7, NCPDP or X12 data, and handling patient data to HIPAA standards.
• Time as a founder, freelancer or early startup engineer.

How We Hire

1. Send us proof. A link to something you shipped that people use, and about 200 words on a problem you found and how you fixed it. A note from someone who used it helps.
2. Written work sample. We send you a real workflow with names removed. You write back where you think it is stuck, what you would build first and what you would leave out. It takes about three hours, and we pay for your time.
3. 90-minute live session. You talk us through your answer. Then we give you a broken agent and its logs, and you work out what is wrong.