Applied Research
Selected Applied Research
Based on A/B tests and product launches.
Role legend: [L] Lead Author · [CL] Co-Lead Author · [CA] Corresponding Author · [C] Contributing Author
Agentic Solutions
- Perspectives on agentic RecSys — under review at ACM ToRS, [L]
- Reducing output variance in multi-agentic evaluation, NeurIPS25-W, [L]
- Multi-agentic evaluation through crowd-sourcing agents, NeurIPS25-W, [C]
- Explanation generation for RecSys, ICML25-W, [CL]
Retrieval, RAG, and Language/Vision Generation
- Text generation with user implicit feedback, NeurIPS25-W, [CA, CL]
- Personalized recommendation with agentic RAG, SIGIR25-W, [L]
- Geometric RAG for layout design, SIGIR25-W, [L]
- Abstractive keyword extraction, IEEE BigData23, [L]
Algorithms
- User Inference and Combinatorial Assortment Optimization, under review at MSOM; also covered in my Ph.D. Thesis, [L]
ML/DL for Industrial IR / RecSys (workshops)
- LLM-based embeddings for recommendation, ICML23-W, [L]
- Deals recommendation based on prospect theory, IEEE BigData22, [L]
- GNN-based similar item recommendation, ICDM23, [C]
- Seller-side fairness in online marketplaces, NeurIPS23-W, [CA]
Tutorials & Talks
- Aug 25: Invited talk on Modern Topics in Recommender Systems at TU Wien, Vienna, Austria
- Sep 25: Tutorial on Agentic RecSys at ACM RecSys25,
- Aril 26: Invited talk on Agentic Retail, Richard A. Chaifetz School of Business, SLU
- July 26: Tutorial on Production Grade Agentic Recommender Systems SIGIR26
Past Upcoming
US Patents
Four additional patents pending.
PhD Thesis
University of Illinois Urbana-Champaign — Urbana, IL, USA Ph.D., 2021
- Thesis: Choice modeling and recommendation optimization in presence of context effects, [Link]
- Advisor: Professor Xin Chen, [Link]
- Honors / Awards: (i) Received full funding from the Walmart Labs Personalization team based on direct applicability of thesis research. (ii) Won the Hansen Fellowship, awarded to bright graduate students.