I am a Research Assistant at the University of Nebraska Omaha, currently working on machine learning and wearable sensor applications in dementia care. I have prior experience with vision-language models and their applications in medical imaging. During my final year of undergraduate studies, I worked as a Research Intern at the Nepal Applied Mathematics and Informatics Institute for Research (NAAMII), where I explored computer vision and multimodal learning for healthcare.
Courses: Advaced Concepts in Programming Languages, Design and Analysis of Algorithms, Software Specification and Design, Advanced Topics in AI
Courses: Linear Algebra, Calculus, Operating Systems, Analysis Of Algorithms, Theory of Computation, Artificial Intelligence, Computer Networking, Database Management System, Data Warehousing and Data Mining
A comparative analysis of fine-tuning strategies for CLIPSeg on abdominal organ segmentation in CT scans, evaluating full fine-tuning, decoder-only tuning, and adapter-based methods focusing on trade-off between Dice score and training efficiency.
Designed and deployed a custom neural network library by implementing key components including forward propagation, back propagation, activation functions and evaluation metrics from scratch using NumPy.
Implemented a custom ClipSeg model from scratch in PyTorch for zero-shot image segmentation.
Conducted a series of experiments to explore and visualize the impact of Batch Normalization when training Convolutional Neural Networks (CNNs).
Developed a document management system for Nepali-language papers by fine-tuning Tesseract OCR for text extraction and training a classifier based on LSTM for automated classification achieving 4 % Character Error Rate (CER) in OCR and F1 score of 89 in classification. Containerized the system using Docker and deployed on AWS EC2.
Python, C, C++, SQL, Bash
FastAPI, MySQL, HTML, CSS
PyTorch, HuggingFace, Snowflake, Airflow
Git, Docker, MLflow, AWS
Research, Writing, Communication, Teamwork