I built an Enterprise Agentic RAG System as an independent project, using LangGraph to route queries, maintain conversation memory, and support structured reasoning. The system uses Qdrant Cloud, Gemini Embeddings, and FlashRank for retrieval, with NeMo Guardrails to filter unsafe or off-topic requests.
On a team project, I developed a multilingual cybersecurity complaint processing system for text, voice, and image inputs in English, Hindi, and Marathi. Its classification engine reached 87% accuracy, and its named entity extraction achieved 92% precision.
I also worked on an AI-based spleen disease detection system using ensemble machine learning, cross-validation, and SHAP visualizations. As Technical Lead of the AI/ML Club at St. Vincent Pallotti College, I organize workshops and mentor students on machine learning projects.
