• Design and build single and multi-agent systems with planning, memory, and tool use • Build and operate MCP servers with secure schemas and permissions • Develop agentic workflows using LangGraph or equivalent frameworks • Integrate LLMs via SDKs; manage prompts, structured outputs, and tool calling • Define and run LLM evaluations for quality, correctness, latency, cost, and regressions • Build observability and reliability infrastructure: logging, tracing, retries, state management • Optimize performance and cost from prototype to production • Mentor engineers and establish agentic best practices