Build the agent runtime and the LLM application layer that powers Agents – the ai-gateway that routes across Anthropic, OpenAI, and our own served models, plus tools, MCP, deep-research orchestration, and the App Builder codegen loop. You will turn frontier and owned models into product features people actually depend on.
Requirements:
2+ years building LLM-powered products in production
Strong TypeScript or Python
Hands-on with agent frameworks, tool-use, and MCP
RAG and vector search (Qdrant, pgvector, or similar)
Rigorous about prompting, evals, and measuring model quality
Bonus: streaming tool calls, multi-agent orchestration, Connect/gRPC
2+ years building LLM-powered products in production
Strong TypeScript or Python
Hands-on with agent frameworks, tool-use, and MCP
RAG and vector search (Qdrant, pgvector, or similar)
Rigorous about prompting, evals, and measuring model quality
Bonus: streaming tool calls, multi-agent orchestration, Connect/gRPC
This position is open to all candidates.
