Senior AI Engineer - Agent Systems

About

This is not a typical AI Engineer position. You'll be part of an AI-native systems team building agent-based applications in production - designing and deploying AI agents that operate in real-world environments, not just prototypes. LLMs are not a feature here — they are the core runtime.

You’ll work on systems where agents plan, reason, use tools, and deliver outcomes end-to-end. We’re looking for engineers who don’t just understand AI — but can turn it into working products.

What you’ll do

• Build and deploy AI agents in production (not demos)
• Create LLM orchestration pipelines and multi-step workflows
• Develop tool-using agents (APIs, functions, external systems)
• Work directly with clients to turn problems into working solutions
• Influence product direction and experiment with new agent capabilities
• Own systems end-to-end: from idea → build → deploy → iterate
• Work extensively with and on internal coding-agent tooling (Claude Code, Codex, skills, hooks, ...)

What we’re looking for

• Strong software engineering background (4+ years or equivalent)
• Hands-on experience building with LLMs in real applications
• Experience with: Coding Agents, harness engineering, APIs and backend systems (Python preferred)
• Ability to build and ship independently

Nice to have

• Research experience + publications
• Evals/Benchmark knowledge
• Experience with computer-use or browser-use agents
• Full-stack capability (you can ship end-to-end)
• Background as a founder, early/founding engineer

What makes you stand out

• You’ve built something real: A product, startup, or widely used tool
• You have public proof of work: GitHub, side projects, demos, or benchmarks
• You think in products and use cases, not just models
• You’re comfortable working directly with users or clients

Why join

• Work on real AI systems, not slideware
• Build and deploy agents that actually get used
• High ownership, low hierarchy
• No line management — this is a pure builder role
• Fast iteration, real-world deployment
• Small team of highly autonomous engineers
• Shape the direction of an emerging agent-first platform
• Flexible setup with Berlin base or full remote option