Senior AI Engineer, LLM Pipelines

The role

You build the multi-agent LLM pipeline at the heart of our decision-intelligence platform, working closely with our CTO. You shape how a question is framed, how it becomes research tasks, and how agents debate the findings and people check them before a report goes out.

Why this role

• You own a whole pipeline, from the first question to the reviewed report, not a single model call.
• You design how AI agents check, debate and correct one another.
• What you build decides whether people can trust the answer.

What you will do in your first year

• Build question intake: agents that help a user sharpen and challenge their question before research starts.
• Turn each question into a plan of research tasks, with independent checks that catch weak plans before they run.
• Turn research results into findings people can rely on: weighing conflicting evidence, structured debate between agents, and review steps where people approve findings or send them back.
• Make sure every answer rests on sources that people have approved.
• Keep model use reliable and within budget across several model providers.

What you bring

• Several years designing and owning production systems in Python. Show us one.
• Hands-on work with LLM pipelines: multi-step orchestration, structured outputs, prompt design, evaluation.
• Workflow engines or job orchestration (Temporal, Airflow, Celery or similar).
• Experience with several model vendors while keeping cost under control.
• A reviewer’s eye: you turn a vague goal into a precise work order an AI coding agent can build from, and you can show us a time you caught an agent’s code or tests being wrong.

Nice to have

Temporal; FastAPI and Postgres; LiteLLM or another model gateway; multi-agent debate or LLM-as-judge evaluation; JSON Schema contracts.