Scholar
Bhuvesh Kumar
Google Scholar ID: fKGi5JwAAAAJ
Snap
Citations & Impact
Citations
88
H-index
5
i10-index
2
Publications
13
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Publications: 'Sequential Data Augmentation for Generative Recommendation' (WSDM 2026), 'Learning Universal User Representations Leveraging Cross-Domain User Intent at Snapchat' (SIGIR 2025), 'Generative Recommendation with Semantic IDs: A Practitioner’s Handbook' (CIKM 2025, Best Paper Award in Resource Papers), 'Private Mechanism Design via Quantile Estimation' (ICLR 2025), 'Revisiting self-attention for cross-domain sequential recommendation' (KDD 2025), 'Accelerated Federated Optimization with Quantization' (IEEE Data Engineering Bulletin 2023), 'ActiveHedge: Hedge meets Active Learning' (ICML 2022, Spotlight), 'Observation Free Attacks on Stochastic Bandits' (NeurIPS 2021), 'Bridging Truthfulness and Corruption-robustness in Multi-Armed Bandit Mechanisms' (Incentives in ML Workshop at ICML 2020), 'Learning Auctions with Robust Incentive Guarantees' (NeurIPS 2019).
Research Experience
- Currently a Research Scientist at Snap Inc., previously a Machine Learning Scientist at TikTok, where he designed several components of TikTok’s recommendation system.
Education
- Ph.D. in Computer Science from Georgia Institute of Technology, advised by Prof. Jacob Abernethy and Prof. Jamie Morgenstern; Bachelor's degree in Computer Science and Engineering from IIT Kanpur.
Background
- Research Interests: Machine learning for large-scale recommendation systems, generative recommendation, and representation learning using multimodal user interaction data. Brief Introduction: Working on the User Modeling and Personalization (UMaP) team at Snap Research, focusing on advancing the state of the art in machine learning to make effective and responsible decisions in complex environments.
Miscellany
- Links to CV, Google Scholar, GitHub, and LinkedIn are provided on his personal website.
Co-authors