Job Description

As the first Staff Engineer focused on Agentic AI, you will own the platform end-to-end, not manage people. You will work on: - Agent orchestration: tool calling, planner/executor and multi-agent patterns, MCP/A2A integrations with enterprise systems - Making enterprise context reliably available at inference time—retrieval, memory, hybrid search, reranking, access controls - LLMOps that isn’t hand-wavy: evals and golden datasets, prompt/agent versioning, tracing, token and cost budgets, CI/CD for agents - Safety in a regulated enterprise: prompt injection, data exfiltration, PII handling, human-in-the-loop, audit logging. You should have: - 7–10+ years building and scaling production systems, with recent, hands-on GenAI/agentic work in production (not just demos) - Shipped agents or LLM pipelines at scale - Solid data chops — vector stores, chunking, hybrid search, knowing when retrieval is the wrong answer - Cloud experience (we’re primarily AWS Bedrock); comfortable with containers and serverless - Track record of technical leadership and mentoring without needing a manager title. Nice to have: LangGraph/Semantic Kernel/PydanticAI, GraphRAG, knowledge graphs, LangFuse/LangSmith, multimodal.