Founding Senior AI Software Engineer

# Founding Senior AI Software Engineer Own an AI-native platform module from problem framing to production, building with Claude Code as a core tool.

Apply by email

Ask a questionTYPEContract to full-timeLOCATIONRemoteAPPLYteam@xeniostechnology.comxenios is building the platform that puts a coach in every client's pocket: available around the clock, sounding like the real coach, and scaling one coach's expertise to serve many more clients without losing the personal touch that makes coaching work. We are a small, fast-moving team shipping real product, not decks, not committees, not endless planning cycles.

Why this matters. A great coach can only serve 20 to 40 clients before quality degrades. Scheduling, follow-ups, check-ins, program adjustments, motivation, and accountability grow with every new client. Clients need their coach early in the morning at the gym, late at night before they order takeout, and in the moments they are about to skip a workout. The coach cannot be there around the clock, but the coach's knowledge, philosophy, voice, and judgment can.

We are building the system that clones that expertise into an AI that acts on the coach's behalf: drafting personalized plans, giving real-time form feedback through the phone camera, analyzing food logs, sending check-ins, encouraging clients, and escalating to the human coach when it matters.

The role. We are looking for senior AI software engineers who want to own an entire platform module, from problem framing through production deployment. You will run Discovery Builds: researching the domain, building proofs of concept, turning working prototypes into specs, and shipping production software. You will also pick up well-scoped Feature Sprints and ship them fast. We follow a build-to-learn process: working prototypes are specs that run.

If your instinct is to demand a detailed requirements document before you open an editor, this is not the right fit. This is not a role where someone assigns you tickets each morning. You get context, alignment on the goal, and then you run with it.## What you will do* →Own a major AI-native platform module end to end: the agent orchestrator, the coach personality system, the client experience, the real-time form-feedback pipeline, or the communication engine.

Architecture, implementation, testing, deployment, iteration.* →Run Discovery Builds: problem brief, proof of concept, domain-expert review, refined spec, production build. Test the hardest 10 percent of the problem first, then hand the proven approach to Claude Code.* →Ship Feature Sprints: take a brief with clear acceptance criteria and ship it in 1 to 3 days using Claude Code.* →Write machine-ready PRDs: working proof-of-concept code, concrete acceptance criteria, data requirements with schemas and API payloads, and a verification plan precise enough for an AI agent to execute against.
* →Collaborate through a generate, review, refine loop with domain experts.* →Communicate asynchronously through learning logs instead of daily standups. If there is no blocker, protect deep work.## Who you are* →You have built entire platforms from scratch, with or without AI tools. You have taken something from zero to production and kept it running.* →You think in specs that machines can execute, and you know a working proof of concept communicates architectural intent better than prose.
* →You have an ownership mindset and you are a strong teammate who makes the people around you more effective.* →You solve the hardest problem first. Here that means security and compliance (HIPAA, SOC
2), the agent guardrail pipeline that keeps AI outputs inside safety boundaries, and the event-driven orchestration layer that processes 15,000-plus tasks per day…