Projects
My project work balances research depth, engineering execution, and practical outcomes.
Developed an AI-powered evaluation model for descriptive answers using BERT, SBERT, OCR, semantic similarity, and a reviewer-facing web interface.
Python, TensorFlow, Hugging Face, OpenCV, Flask, Streamlit
Researching hybrid architectures for stronger language modeling efficiency in low-resource environments.
NLP, model optimization, efficiency research
Built a multi-label classification model to identify mental health conditions from textual user data with attention-based NLP methods.
BERT, attention mechanisms, text classification
Developed a YOLOv8-based safety pipeline for vehicle detection, speed estimation, collision risk prediction, and live alerting.
YOLOv8, object tracking, safety analytics
Created an LLM-enabled tool for dataset processing, intelligent web search, and targeted information extraction through a user-friendly interface.
LLMs, search, structured extraction
Designed a vision-based deep learning system to identify hazards, obstacles, and lane issues for safer two-wheeler mobility.
Computer vision, road safety, deep learning
Built a CNN-based recognition workflow using OpenCV and FaceNet embeddings for identity verification.
CNN, OpenCV, FaceNet