Computer Vision Foundations
Learning visual representations that remain stable under domain shifts, sparse labels, and practical deployment constraints.
Research Areas
My current work explores how AI can build stronger visual intelligence systems with better generalization, reliability, and multimodal understanding.
Visual Anomaly Detection
Detecting rare patterns and out-of-distribution behavior in visual data for quality control and scientific use-cases.
Self-supervised Representation Learning
Designing pretraining strategies that reduce dependency on expensive annotation and improve transfer across tasks.
Generative and Multimodal Models
Integrating generative modeling and vision-language techniques to support analysis, retrieval, and robust reasoning.