Built a full-stack AI agriculture platform with a RAG pipeline (GPT-4o, LangChain, vector embeddings) at 90%+ relevance and sub-500ms responses, a YOLOv8/PyTorch disease classifier at 85% accuracy across 20k+ images, and a Spring Boot REST API deployed on AWS with GitHub Actions CI/CD.

Built a secure enterprise RAG chatbot (FastAPI, LangChain, LangGraph, FAISS) for real-time PDF Q&A, cutting document search time by 60%. Integrated a Scikit-learn query classifier with RBAC and audit logging, blocking 92% of unauthorized access attempts. Deployed LLM services on Azure OpenAI and AWS Bedrock, using Databricks to orchestrate data pipelines and benchmark model quality.

Developed and integrated 15+ REST APIs into a fraud detection platform, provisioned ML model artifact infrastructure on AWS achieving a 30% reduction in data retrieval time, and implemented secure JWT-based authentication with sub-500ms OTP delivery.

Optimized Spring Boot batches with multithreading (25% faster throughput) and shipped reusable React + REST components that cut reporting bugs by 50%.