Conducted research on interpretability and control in neural networks, analyzing how internal representations vary across task demands.
Evaluated semantic representations using SVM probing, dimensionality reduction, pairwise distance analysis, and multidimensional scaling (MDS).
Software Engineer Intern
Viettel Group
Designed and containerized AI model evaluation environments using Docker, standardizing workflows and ensuring reproducibility across platforms.
Implemented local LLM inference pipelines via Ollama, enabling rapid model testing and benchmarking without dedicated GPU infrastructure.
Presented technical findings to senior engineers, including projected productivity gains, resource requirements, and deployment recommendations.
Optimized Viettel’s AI Tessel computer vision pipeline for multi-angle grocery shelf image analysis, increasing matching accuracy by 30% and reducing end-to-end processing time by 25%
Undergraduate Research Assistant
University of Utah, School of Computing
Developed a novel LSTM-based model to predict tendon robot shapes from current configurations, surpassing prior state-of-the-art performance by 21%.
Engineered a Bayesian optimization framework to automate surgical task execution, improving tissue retraction efficiency by 27% and attachment point detection accuracy by 15%.
Contributed to lab publications and presentations, demonstrating applicability of ML techniques in robotic surgical automation.
Education
M.S. in Computer Science
University of Maryland - College Park
B.S. in Computer Science; Minor in Statistics, Mathematics
University of Wisconsin - Madison
GPA: 3.7/4.0
Relevant courses: Algorithms, Data Structures, Statistical Modeling, Object-Oriented Programming, Operating Systems, Database Systems, Probability Theory, Linear Algebra, Optimization, Deep Learning, Interpretable ML