Hello, I'm
Surya Vasanth.

Built an end-to-end Intelligent Operations AI Assistant to eliminate decision latency on the factory floor. Using LangChain and ChromaDB, I developed a RAG-based retrieval system that lets plant managers instantly query SOPs, historical maintenance logs, and downtime reports in plain language. Integrated live OEE metrics via a Streamlit dashboard, combining real-time performance data with contextual technical retrieval into a single unified interface. A functional prototype designed to prove that Generative AI can deliver measurable, domain-specific value in industrial environments.

Performed in-depth EDA and feature engineering to identify key ridership patterns and developed a GRU-based deep learning model in PyTorch to forecast daily Citi Bike demand. This project successfully captured complex temporal and weather-driven dependencies to provide actionable insights for fleet rebalancing and operational planning.

Designed a NLP pipeline using NLTK and Scikit-learn to process and classify real-world disaster tweets from noisy social media data. Achieved 85.9% test accuracy, demonstrating a highly effective and reliable tool for filtering critical information for emergency response teams.