Research

I am interested in the area of representation learning. More recently, I have developed an interest in diffusion models (subjects include alignment, test-time scaling, controlled generation).

Please refer to my Google Scholar profile for details on the publications I have been a part of.

Conference tutorials

• Practical Adversarial Robustness in Deep Learning: Problems and Solutions (CVPR’21)
• Foundational Robustness of Foundation Models (NeurIPS’22)
• All Things ViTs: Understanding and Interpreting Attention in Vision (CVPR’23)

Invited talks, demos, etc.

• SoTA Diffusion Models with 🧨 diffusers (slides and recording)
  • IBM Research (October 17, 2023)
  • The Dyson Robotics Lab, Imperial College London
  • Department of Statistics, University of Oxford
• 🧨 diffusers for research at VAL, Indian Institute of Science (IISc), June 12, 2023. Slides are here.
• Controlling Text-to-Image Diffusion Models: Assorted Approaches
  • BigMAC ICCV Workshop in 2023 (slides, recording)
  • UCL in June 2024
  • Texas A&M University in September 2024 (slides)
  • UC Berkeley in October 2024 (slides, recording)
  • ICDMAI’25 (Slides)
• Demo of 🧨 diffusers at ICCV 2023 (Tweet).
• A talk on diffusion models, ETH Zurich (May 06, 2024). Slides are here.
• Transformers in Diffusion Models for Image Generation and Beyond, CS25 v5, Stanford (May 27, 2025). Slides are here.

For regular talks, refer here.

Teaching assistance

Served as a TA for Full Stack Deep Learning’s 2022 cohort.

Reviewing

• Conferences: ICCV’25, ICML’25, CVPR’25, ICLR’25, NeurIPS’24, AAAI’23, ICASSP’21 (sub-reviewer)
• Workshops: UDL workshop (ICML’21)
• Journals: TMLR, Artificial Intelligence (Elsevier), IEEE Access.

Misc

• Released a dataset for large-scale multi-label text classification (joint work with Soumik Rakshit).
• Exploration on instruction-tuning Stable Diffusion.