Vector Institute, Toronto, Canada

• Researching deepfake video detection systems for manipulated multimodal content.
• Building training and validation recipes to improve robustness in real-world forgery scenarios.

University of British Columbia, Vancouver, Canada

• Investigating emergent modularities in foundation models and brain-like functional specialization.
• Proposed FLoRG: federated fine-tuning with low-rank Gram matrices and Procrustes alignment.
• Proposed ZorBA: zeroth-order federated fine-tuning with heterogeneous block activation.
• Developed ASFL with online resource allocation and improved learning efficiency.
• Designed a GNN-based framework for decentralized federated learning over D2D networks.