Prathyusha Chatla

I recently built a multi-agent healthcare query system: a three-agent CrewAI pipeline backed by a RAG knowledge base in ChromaDB, LangChain session memory, and Pydantic-validated outputs.

The system includes PII masking, prompt injection detection, and groundedness checks, plus an Autogen review stage and an LLM-as-judge evaluation framework. I deployed it as a FastAPI service with structured logging and response caching.

I believe in evidence-driven development, validating design decisions through measured results. I’m open to Agentic AI Engineering roles where deep system-building meets real-world impact.