news

Sep 02, 2026 Excited to share our latest work, It Takes Two to Match: Co-Evolving Generative Retriever with Reinforcement Learning. We introduce CoGR, a generative retrieval framework in which LLMs generate matching keywords for both queries and items, and the two sides co-evolve through reinforcement learning. Check out my posts on LinkedIn and X.
Jun 28, 2026 I will be attending ACL 2026 in San Diego. Excited to meet with friends there!
May 02, 2026 I’m happy to share that I joined Apple as a Senior Machine Learning Engineer this January, working on LLM post-training, RL policy optimization, and alignment evaluation for large-scale retrieval. I look forward to pushing the boundaries of reasoning-driven retrieval and alignment. Read more on LinkedIn.
Apr 06, 2025 Excited to share that our paper, ColBERT-serve: Efficient Multi-Stage Memory-Mapped Scoring, was accepted and presented at the 47th European Conference on Information Retrieval (ECIR 2025). The work introduces a memory-mapped multi-stage scoring approach that makes late-interaction neural retrieval (ColBERT) far more efficient and scalable for serving LLM-powered retrieval systems.