My work broadly lies in Natural Language Processing (NLP) and is strongly motivated by applications of Large Language Models (LLMs) to enterprise settings. In particular, I focus on modeling users’ intents and preferences—what NLP often refers to as pragmatics—and on designing the system-level apparatus needed for LLMs to remain pragmatically grounded when deployed in complex, error-prone real-world environments.

I have pursued this agenda from multiple complementary angles. This includes studying mechanisms for planning and reasoning in the large, ambiguous action spaces typical of enterprise settings (e.g., ICML 2026, ACL 2024, ICLR 2024, EMNLP 2021); analyzing how dataset characteristics affect LLMs’ capacity to perform expert-level tasks (e.g., EMNLP 2025, VIS 2024, EMNLP 2022); and examining the extent to which the behaviors learned by LLMs faithfully capture those of domain experts (e.g., CHI 2025, VIS 2024, COLING 2025).

Looking forward, I am particularly excited about the design of agent-based systems that support the workflows of teams rather than individual users (e.g., ACL 2026 or here at Adobe Research’s blog). This includes scenarios in which multiple stakeholders collaborate toward a shared goal—such as designers, art directors, copywriters, and brand managers working together on campaign briefs—moving beyond the conventional single-user paradigm that dominates current AI systems.

Beyond publications, my research has shipped into production features in two flagship Adobe Experience Cloud products, now in use by multiple Fortune 100 companies. I contributed to the design and deployment of custom LLMs for question answering in AEP AI Assistant (covered by granted patents: US Patent App. 18/486,603; 18/485,204; 18/504,256; 18/589,065; 18/612,566) and for automated insight generation in CJA Intelligent Captions (covered by granted patents: US Patent App. 18/338,033; 18/321,602), considered Adobe Experience Cloud’s first GenAI feature when it went live in 2023. Overall, I have co-authored more than 30 patent filings on novel AI methods since 2015.

Prospective Interns

In addition to collaborating with outstanding full-time researchers and stakeholders on the design of these methods, I have also had the privilege of mentoring interns who were true rising stars. Here is a selection of them:

1. Lin Ai (2025 intern, from Columbia University)
2. Jiale Liu (2025 intern, from Pennsylvania State University)
3. Neha Srikanth (2024 intern, from University of Maryland)
4. Minsoo Kim (2023 intern, from Seoul National University)

If you have published work in competitive venues that aligns with what we do at Adobe Research and are considering applying for a summer internship, I’d be happy to hear from you. Feel free to reach out at {first_name}.{second_last_name}@adobe.com and share a short description of your papers.