Gokulan M

AI engineer specializing in computer vision and sequence-based deep learning systems using PyTorch and TensorFlow. Developed ResNet architectures, trained YOLO models for real-time object tracking, and created anomaly detection pipelines with LSTM and VAE. Built RAG-based NLP applications utilizing LangChain, FAISS, and Hugging Face for large document collections. Proficient in end-to-end model training, evaluation, and deployment, with a focus on research-driven challenges.

MedSnap – AI-powered medical research platform

• Built an AI tool to summarize research abstracts, skim full papers, and track medical news for students and professionals
• Implemented NLP models (DistilBERT, BART, T5), and deployed an interactive Gradio-based web application on Hugging Face Spaces for real-time insights

Human AI crisis tracker

• Developed an AI-driven dashboard to detect high-risk mental health signals from Reddit data
• Applied NLP models (DistilBERT, VADER) for sentiment and contextual analysis, and used geospatial analytics (Folium, GeoPy) to visualize crisis trends by location

AI-powered mining regulatory chatbot

• Built a RAG-based chatbot for the mining industry using 52+ official acts, rules, and DGMS circulars as the knowledge base
• Integrated secure authentication, accident case analysis, and multilingual (Hindi–English) support, deployable on the web

• 1st Place, Techgyan Hackathon (IITM), Gen AI RAG Application
• 3rd Prize, COPS CrtI+Shift+Intelligence Hackathon, MedSnap AI
• Certified in MLOps, Duke University, AWS SageMaker, Azure ML