Arvind Ramesh

I've built AI systems including CareChat, an automated mental health screening platform that combines LLM agents, hybrid retrieval, and specialized ML modules. Its evaluation framework achieved 91% diagnostic accuracy versus a 78% baseline across 200+ test cases.

My work spans end-to-end ML workflows, from data preprocessing and EDA to model benchmarking, validation, deployment, and reporting. For predictive maintenance using NASA Turbojet Engine data, I benchmarked five models and published research at SMARTGENCON-2025.

I work with Python, LangChain, RAG pipelines, semantic search, NLP, Flask APIs, SQL, and data visualization tools to turn structured and unstructured data into reliable analytical systems.