At Globallogic India, I created an end-to-end multi-agent automation system that extracts project information, generates outputs, and delivers them back to source tools.
I developed processing agents using GPT-4, Gemini Pro, Llama 3.3:latest, Bedrock, and Azure OpenAI models for code generation and text summarization. I also implemented RAG with a Neo4j graph database for contextual retrieval.
I designed a supervisor agent to orchestrate workflows and implemented a production-ready Cursor MCP command for JIRA and Azure DevOps ticket context. I deployed FastAPI microservices with Docker.
My work achieved 74% accuracy in generative output classification and validation, accelerated development by 30%, and improved code quality by 25%. I also built machine learning projects in facial recognition, sentiment analysis, and disease prediction.
