Senior AI Engineer, Agents and Platform (f/m/d)

Role overview

A senior engineering position focused on the platform that powers multi-agent AI products for the commercial side of professional sport. The mission is to convert institutional knowledge held by sponsorship, partnership, and pricing experts into finished deliverables such as ROI reports, pitch narratives, pricing analyses, and partner fit assessments. The central engineering challenge is moving beyond impressive demos toward agent systems that are dependable at scale, with correct outputs, strict tenant isolation, and outputs trustworthy enough for real commercial decisions.

Responsibilities

- Own the architecture of the agent harness, covering tool orchestration, durable long-running workflows with resume after failure, memory and context management, retrieval, structured output, model routing, fallbacks, and rollback - Design the tool layer and model-context interfaces, including versioning, per-customer scoping, and structured logging of every agent action - Partner with infrastructure and security leadership to enforce tenant data isolation, least-privilege tool access, prompt-injection defenses, and complete audit trails - Build evaluation pipelines where every change is tested against practitioner-authored evaluation sets, with traces, costs, and timing visible per run - Lead a small senior engineering team, set engineering standards, run code reviews assisted by coding agents, and ship releases that can roll back in minutes - Decide what to build, buy, or retire across the platform and document those decisions so people and agents work from a shared source of truth

Requirements

- Eight or more years building and operating production software, with deep backend experience across APIs, relational databases, queues, and durable workflows - A track record of shipping language-model products to real users, with the ability to describe a failure and the changes made afterward - Hands-on experience with retrieval systems and production agents beyond prototypes, plus the judgment to recognize when an agent is the wrong tool - A rigorous approach to evaluations, treating passing tests with skepticism until failure modes are understood - Daily fluency working alongside multiple coding agents while keeping code legible and owned by humans - A security mindset developed from running systems that held customer data - A demonstrable project or product that can be reviewed - Professional-level English; German is a plus

Nice to have

- Experience designing vendor-agnostic model abstractions or multi-provider routing - Background operating multi-tenant SaaS or other regulated environments

Benefits and work setup

- Platform ownership for a new product with paying customers from the outset - Small senior team, minimal bureaucracy, and direct customer proximity - Choice of hardware and personal tooling - Competitive compensation reflecting the seniority and scope of the role - Startup pace combined with the stability and network of a larger organization - Work that supports commercial outcomes across clubs, leagues, and federations