Ramana Gangarao
Ai Engineer @Optum
Signup · Get unlimited contacts
WORK HISTORY
Ai Engineer @Optum
San Francisco, CA, US
As an AI Engineer at Optum, I build intelligent systems for document understanding and workflow automation, contributing directly to ContractAI and customer care log classification. I design ingestion pipelines using vector and graph storage, leveraging Neo4j to build knowledge graphs that model contractual entities and semantic relationships. To enable multi-step reasoning, I implement parallel agent execution, ReAct loops, and supervisor patterns, while managing agent memory using Google ADK. I process unstructured documents with Google Document AI and expose graph-aware tools through custom MCP servers for agent-driven retrieval. My work spans full-stack delivery using FastAPI and Lovable.dev, ensuring system reliability via DeepEval evaluation workflows, and supporting Google Cloud MLOps by fine-tuning BERT and RoBERTa models on Vertex AI.Tech Stack: Python, Google ADK, Neo4j, MCP, Google Document AI, Vertex AI, FastAPI, Lovable.dev, DeepEval, Google Cloud
EDUCATION
Atlantis University
Master's degree, Information Technology
Keshav memorial institute of technology
Bachelor of Technology (BTech), Electronics and Instrumentation Engineering
Smt B Seetha Polytechnic
Diploma of Education, Applied Electronics and Instrumentation Engineering
ABOUT RAMANA GANGARAO
I am a Senior AI Engineer and Full-Stack Developer dedicated to architecting production-ready intelligence systems.Core Technical Strengths:Agentic AI: Implementing multi-agent workflows using LangGraph, ReAct loops, and supervisor patterns where agents handle specific tasks using planner-executor models. RAG & Retrieval: Building advanced search pipelines including Graph RAG (Neo4j) and vector storage (Milvus) with semantic chunking and cross-encoder reranking for elite accuracy. LLM Ops: Establishing end-to-end observability with Langfuse and setting up automated evaluation pipelines (DeepEval) to mitigate hallucinations and ensure system reliability. MLOps & Engineering: Deep experience in real-time model serving (SageMaker), feature pipelines, and automated CI/CD workflows from my tenure at Wipro. I am passionate about applying Generative AI to solve real-world problems and building AI systems that are not just experimental, but production-grade, secure, and compliant.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.