Safal Thapaliya
PhD student, UConn
LLMs, GNNs, AI for healthcare.
bio
I’m a second-year PhD student in Computer Science and Engineering at the University of Connecticut. My current research focuses on graph learning and combining large language models with graph neural networks. I am particularly interested in building methods that work well without large amounts of labeled data or compute. This focus comes from my background in Nepal, where I worked at NAAMII building healthcare AI prototypes and vision-language segmentation models for low-resource medical settings.
I love to play football (not the american one) and watch any sports that I find interesting.
Aug, 2026
Where LLM Annotators Fail: Label-Free Learning on Graphs with LLMs accepted at EMNLP Findings 2026.
June, 2026
Attended the Machine Learning Summer School 2026 at Columbia University.
Jan, 2025
Dec, 2024
2022–2024
Research assistant at NAAMII’s TOGAI lab under Dr. Bishesh Khanal. Worked on vision-language models, medical image segmentation, and object detection; shipped healthcare AI prototypes alongside KIAS and Dr. Taman Upadhaya.
2022
Solutions Intern at Logpoint on Linux server administration and distributed log handling.
2019–2022
Founding developer at Clamphook.
2018–2022
B.E. in Computer Engineering from IOE, Pulchowk Campus.
publications
* indicates equal contribution
Where LLM Annotators Fail: Label-Free Learning on Graphs with LLMs
EMNLP Findings 2026
Semantic Refinement with LLMs for Graph Representations
arXiv pre-print 2025
Exploring Transfer Learning in Medical Image Segmentation using VLMs
MIDL 2024 (Oral)
Deep-learning assisted detection and quantification of (oo)cysts of Giardia and Cryptosporidium on smartphone microscopy images
Machine Learning for Biomedical Imaging 2024
For a complete list of publications, see my Google Scholar.