• Developed a novel explanation model for LLM-based clustering using Wikipedia concepts. Improved explanation accuracy by 8% to 34% over baseline methods.• Constructed a user-friendly web-based tool to visualize the generated explanations and conduct a user study. Users found the novel explanations more logical, meaningful, and comprehensive.• Conducted hypothesis and statistical analysis, including p-tests and t-tests, revealing statistically significant results that validated the effectiveness of the proposed model.

• Facilitated laboratory sessions to enhance students' understanding of machine learning principles, resulting in improved final exam scores. • Collaborated with faculty in the creation and evaluation of assignments and quizzes.

• Implemented an anomaly detection system for online exams using the VGG-16 model. Managed EDA, finetuning, and AWS EC2 deployment. Reducing human resources required for invigilation.• Produced an online proctoring system with NodeJS and MongoDB. Optimized multitenancy, cutting hardware and operating costs.• Deployed proctoring software on AWS, leveraging Docker for seamless containerization, Load Balancers for efficient traffic distribution, and EC2 instances to ensure scalable, reliable performance within budget.

• Constructed a GMDH neural network model to predict sand's thermal conductivity. Led EDA, training, and evaluation, enhancing prediction accuracy and efficiency.• Presented model findings at a conference, crafting visualizations for diverse audiences, bolstering researcher trust in the model