| Aug 31, 2026 | Cross-Modal Grounding for Calibrated Confidence in Vision Language Models has been accepted at EMNLP 2026 GroundLM Workshop (Conference on Empirical Methods in Natural Language Processing). See you in Budapest, Hungary  |
| Aug 31, 2026 | Attending ACM AI Leadership Summit 2026 in Atlanta, Georgia from August 31 to September 2, 2026. Looking forward to connecting with fellow researchers and industry leaders in AI! |
| Aug 30, 2026 | Echoes of Human Malice in Agents: Benchmarking LLMs for Multi-Turn Online Harassment Attacks has been accepted at ICWSM 2027 (AAAI International Conference on Web and Social Media). See you in Edinburgh, Scotland |
| Jun 15, 2026 | From Actions to Understanding: Conformal Interpretability of Temporal Concepts in LLM Agents has been accepted at the Mechanistic Interpretability Workshop, ICML 2026 in Seoul, South Korea. |
| May 26, 2026 | Excited to share that our tutorial, “Knowledge-Infused Multimodal Learning,” has been accepted at ICWSM ’26, the 20th International AAAI Conference on Web and Social Media! I will be presenting at the University of Southern California, USC Information Sciences Institute, Los Angeles, alongside Agnik Saha and Professor Ugur Kursuncu. We will discuss some of our lab’s recent work on knowledge graph construction and how to design vision-language, knowledge-guided frameworks for more reliable and interpretable multimodal AI systems. Tutorial website |
| May 15, 2026 | Excited to be spending the summer in Seattle, Washington, for my second stint at Siemens Data & AI Lab. I’m looking forward to collaborating on Physical AI initiatives to help design and develop the next generation of AI systems. If you’re in the Seattle area, please feel free to reach out |
| Apr 01, 2026 | New preprint: From Actions to Understanding: Conformal Interpretability of Temporal Concepts in LLM Agents, on step-wise conformal probes for early failure detection in LLM agents.  |
| Feb 01, 2026 | Playing Devil’s Advocate: Unmasking Toxicity and Vulnerabilities in Large Vision-Language Models has been accepted at ACM Transactions on Intelligent Systems and Technology (TIST). |
| Nov 27, 2025 | New preprint with collaborators at KAIST and Amazon: Co-Evolving Agents: Learning from Failures as Hard Negatives. |
| Oct 16, 2025 | New preprint: Echoes of Human Malice in Agents, a benchmark for multi-turn online harassment attacks on LLM agents. |
| May 15, 2025 | Started as an Applied Scientist Intern at Siemens GenAI R&D in Seattle, on the Data & AI Research team, working on multimodal industrial foundation models. |
| May 15, 2025 | Just KIDDIN’: Knowledge Infusion and Distillation for Detection of INdecent Memes has been accepted to ACL 2025 Findings (acceptance rate 19.1%). Grateful for the ACL Student Travel Award. |
| Dec 15, 2024 | Enhancing Cross-Modal Contextual Congruence for Crowdfunding Success was presented at IEEE BigData 2024 (acceptance rate 18.7%), supported by the IEEE BigData Student Travel Award. |