Dharmraj Patil

At EdTech and IIT Bombay, I fine-tuned BERT on the IEMOCAP dataset, achieving 87% accuracy across eight emotion classes. I collaborated with a five-member team to develop transformer-based NLP solutions using Hugging Face.

For my Intelligent Aviation Reliability and Delay Prediction Platform, I architected an end-to-end flight delay prediction platform using Python, SQL, AWS, and Power BI. I implemented AWS-based data pipelines and developed a Random Forest classifier for delay prediction.

I developed AIVA, a multilingual voice assistant that reduced admission query response time by 70%. Its FAISS-based RAG pipeline retrieved answers from more than 500 PDF documents, and its OpenAI API conversational AI supported five languages.

For the Grapevine Precision Health Monitoring System, I engineered a CNN-based disease detection system and trained a VGG16 model to 94% accuracy on grape leaf images. I also developed a React web app for image-based disease diagnosis.