Research Scientist at Meta. PhD in Electrical & Computer Engineering, Purdue University (2025). I work on the alignment, privacy-preserving training, and trustworthiness of large language and foundation models.
LLM & foundation-model alignment (including test-time alignment) · privacy-preserving generative AI · trustworthy and robust machine learning.
📚 Google Scholar · 📄 Full publication list
- TARo: Token-level Adaptive Routing for LLM Test-time Alignment — ACL Findings 2026
- BalancedDPO: Adaptive Multi-Metric Alignment — TMLR 2026
- PrivateEdit: A Privacy-Preserving Pipeline for Face-Centric Generative Image Editing — IEEE TAI 2026
- Domain Adaptive Few-Shot Open-Set Learning — ICCV 2023
- Multi-task Hierarchical Adversarial Inverse Reinforcement Learning — ICML 2023
- Reinforced Sequential Decision-Making for Sepsis Treatment (PosNegDM) — IEEE JBHI 2024
- Multi-source Open-set Deep Adversarial Domain Adaptation — ECCV 2020
- Multiple Instance Learning for Breast-Cancer Histopathology — official code; image classification with weakly-supervised localization
- PosNegDM — official code for the IEEE JBHI sepsis-treatment framework
- Interactive-GradCAM — plug-and-play Grad-CAM visualization notebooks (COCO / ImageNet)
- GPU-Benchmarking — benchmark GPU throughput across model architectures
📈 Activity and contribution graph below.