As a Senior AI Engineer, you'll own the design and reliability of the LLM agents at the heart of our AI-powered platform. This is a hands-on individual-contributor role for someone who has built agents that run in production — and knows what it takes to keep them fast, correct, and observable at scale.
Your work will include:
- Designing, building, and hardening LLM agents with real tool use and multi-step orchestration for data quality, catalog, and governance tasks.
- Owning agent reliability end to end — latency, correctness, and cost — including evaluation harnesses and production observability.
- Designing and evolving our agent tooling layer, including MCP servers and protocol-level integrations, so capabilities can be exposed cleanly and securely.
- Building and tuning RAG pipelines and grounding strategies where agents need reliable access to customer data and metadata.
- Working across our LLM gateway (LiteLLM proxy) and model providers to route, monitor, and optimize inference.
- Setting technical standards for how we build agents, and raising the bar for the whole team through design reviews and pragmatic architectural trade-offs.
What we're looking for
- Strong, modern Python and a track record of shipping LLM agents to production — not just prototypes or demos.
- Hands-on experience with agent architectures: tool calling, orchestration, and the failure modes that come with them.
- Practical experience with MCP (or comparable protocol/tooling design), and the judgment to know when a protocol layer is worth it.
- Experience building agent evaluation and observability — you measure quality and latency, you don't guess at them.
- Familiarity with RAG, retrieval, and grounding techniques and their real-world limitations.
- Comfort with LLM gateways/proxies (e.g. LiteLLM) and provider APIs (Azure AI Foundry, OpenAI, or similar).
- Healthy skepticism toward hype — you surface trade-offs before recommending a direction.
Nice to have
- Kubernetes, AWS, or Azure experience.
- Background in data management, data quality, or governance.
- Exposure to OAuth2/SSO and multi-tenant service design.