You'd be the most senior technical person in the company, in charge of setting the technical direction of the company and owning delivery end-to-end. **What you'll own** You'll build on top of a real production stack: Python/FastAPI on Postgres and Supabase, a SwiftUI iOS app, React and TanStack across our web and admin surfaces, AWS managed with Terraform. Every architectural decision from here is yours and you'll have a say in who works alongside you. - **The data model.** Animal health data arrives fragmented, inconsistently coded, and from sources that disagree with each other. You'll design the model that turns it into one coherent record per animal that holds up as the volume grows and the questions we ask of it get harder. - **AI that has to be right about somebody's dog.** We have LLM features in production today, used by real pet parents asking real questions about their animals. Making them good is a systems problem, not a prompting problem: retrieval that surfaces the right evidence, an eval harness with a golden set and regression gates in CI, a failure taxonomy derived from reading actual traces, a clean handoff to a human when needed. - **Integrations with systems that don't want to be integrated with.** Health data sources with undocumented APIs, unreliable uptime, and payloads that contradict themselves. Idempotency, reconciliation, and a trust layer over data you don't control. - **The security and privacy bar.** Pet health data isn't HIPAA-covered but still gets scrutinized as though it were, so you'll set a bar well above what the law strictly requires and own the work that proves it under real diligence. - **The team after you.** Engineering culture, hiring bar, and the first engineers you bring in. **Your first 90 days** Ship an improvement to the AI experience in week one; our evals run in CI, so you'll know by Friday whether it worked. By day 30, have a new data source live end to end: raw payload to something a pet parent actually reads. By day 90, have built the first version of the learning loop for our longevity model: what we capture, what we learn from it, and how we prove it made the model better. **You're a fit if** Roughly 4-6 years of production engineering. The number is a proxy for the following; if you've done these in three years or in eight, apply. - You've owned a backend service end-to-end in production - You've shipped an LLM feature and then made it measurably better - You've integrated a hostile external system - You're deep in Python and have opinions about async - You've made an architectural decision that other engineers then built on top of - You go from a vague problem statement to a working system without a spec - You actually care about animals. No need to have one yourself, but the passion helps you fit in :) **Bonus:** multi-agent systems in production, a background in healthcare, comfort shipping iOS or React when the product needs it. **You're probably not a fit if** - You need a written spec and a defined scope to start. - Your LLM experience is calling an API inside a request handler. If you've owned retrieval quality as a metric, it isn't. - You want to work fully remotely. We generally value commitment over procedures, but we'd love to make your physical presence a part of Omi Health.