Projects
Early Detection and Severity Assessment of Osteoarthritis
Developed a deep learning model to classify knee osteoarthritis severity levels using ResNet18 and MobileNetV2 architectures.
XGBoost-Gemini LLM Chain for Severity Assessment
Developed a predictive model coupled with an LLM for pregnancy risk management and severity assessment.
Identifying and mitigating Bias in Recidivism Prediction using Responsible AI
Analyzed and addressed biases in recidivism prediction models using responsible AI techniques.
Real Time object Detection in CHUM Hospital Construction Site
Implemented a YOLOv8-based model for real-time identification of construction equipment.
Enhancing Clustering with Deep Learning
Integrated K-means with autoencoders for data insights and interactive clustering refinement.
Realtime Sign Language Recognition
Developed a real-time system to translate hand gestures into text for improved communication.
Emopal - Real-time Emotion Analysis
Created a platform for real-time emotion detection from images, groups, and audio input.
Semantic Textual Equivalence Modeling
Developed a model to determine semantic similarity between text passages.
Constructing a Robust Retrieval-Augmented Generation Framework
Developed a RAG system combining information retrieval and generative models for better text generation.
Shoplifting Detection
Implemented a YOLO-based system to detect and track suspicious behavior in retail environments.