Building MyLow Companion — Lowe's enterprise multi-agent AI platform deployed across ~1,900 US stores, handling 50,000+ daily queries, and designing MCP servers with sub-second memory retrieval at scale.
Multi-agent AI systems, RAG pipelines, MCP server design, and open-source agentic frameworks — especially projects pushing the boundaries of LLM orchestration and enterprise-scale AI infrastructure.
Advancing vector retrieval performance at high concurrency, exploring emerging agentic frameworks, and connecting with AI engineers tackling large-scale production LLM deployment challenges.
Google ADK, LangGraph, and the latest agentic orchestration patterns — deepening expertise in FastMCP server architecture, advanced prompt engineering, and scalable async AI pipelines.
Building multi-agent AI platforms from scratch, designing RAG pipelines and MCP servers, scaling LLMs to enterprise production (50K+ daily queries), vector database optimization (Milvus, Weaviate), and driving 90%+ cost reduction with GenAI automation.
I architected Lowe's flagship AI platform before LangGraph or Google ADK even existed — building the entire multi-agent orchestration layer from scratch, now live across every US Lowe's store.
