Nov 2022 – Jan 2026 · 3 yrs 2 mos · Greater Seattle Area

• LLM/SLM: (post-training, reinforcement learning, alignment), Agentic Applications, LLMs for entity resolution. These two publications capture some of my work:
• ICLR 2026: Self-Aligned Reinforcement Learning for Reasoning Models - https://arxiv.org/pdf/2509.05489
• EMNLP 2024: Training LLMs for Entity Resolution - https://aclanthology.org/2024.emnlp-main.352/

Jun 2021 – Nov 2022 · 1 yr 5 mos · Mountain View, California, United States

• Primary contributor to neural search, retrieval, and content understanding for Microsoft365 Designer - https://designer.microsoft.com/. Lead author/contributor to 6 granted patents:
• https://patents.google.com/patent/US12326867B2/en [Domain specific content retrieval]
• https://patents.google.com/patent/US12314347B2/en [Multimodal retrieval]
• https://patents.google.com/patent/US12307750B2/en [Knowledge Distillation]
• https://patents.google.com/patent/US12242491B2/en [Personalized search]
• https://patents.google.com/patent/US12045279B2/en [Vision based retrieval/search]
• https://patents.google.com/patent/US11841911B2/en [Text based retrieval/search]
• Focus areas: VLM/LLM-based Content Analysis, Multimodal Information Retrieval, Zero-Shot Learning, Knowledge Distillation.

May 2020 – Jul 2020 · 2 mos · San Francisco Bay Area

• Scalable deep multi-task knowledge graph embeddings for use in diverse customer prediction and regression tasks at Visa.
• Patent: https://patents.google.com/patent/US20220114456A1/
• Publication: https://link.springer.com/chapter/10.1007/978-3-031-05933-9_21

May 2019 – Aug 2019 · 3 mos · San Francisco Bay Area

• Transferable neural models for cross-domain recommendation with contextual invariants, trained on data-rich users and transferred to sparser domains.
• Patent: https://patents.google.com/patent/US20210110306A1/
• Publication: https://dl.acm.org/doi/abs/10.1145/3397271.3401078

Jan 2016 – Jan 2021 · 5 yrs · Urbana-Champaign Area

• My research develops robust deep-learning models and training methods under real-world data constraints for several applications - Information Retrieval and Ranking/Recommendation, Graph Mining, User Behavior Modeling, and NLP. My work was published at multiple flagship conferences - ACM SIGIR, ACM CIKM, ACM WSDM, ACM Webconf, IEEE ICDM, ACM RecSys - spanning data mining and machine learning for Information Retrieval, Recommender Systems, Graph Mining, NLP, and User Modeling.
• Google Scholar - https://scholar.google.com/citations?user=pWaAozwAAAAJ

May 2015 – Jul 2015 · 2 mos · Greater Bengaluru Area

• Scalable and Diversified Query Expansion by exploiting the Wikipedia entity network with a generalizable graph-based semantic formulation.
• Publications:
• https://link.springer.com/article/10.1007/s11280-017-0468-7
• https://link.springer.com/chapter/10.1007/978-3-319-48740-3_11

May 2014 – Jul 2014 · 2 mos · Greater Bengaluru Area

• Episodic pattern mining in conversion analysis for marketing campaigns and e-commerce applications.