news

Aug 20, 2026 Forget Without Compromise is accepted to EMNLP 2026 — streaming KV-cache eviction under fixed budgets.
May 18, 2026 Joining the Vertex Gemini Inference team as a Research Intern this summer.
May 01, 2026 FAFO and Scout Before You Attend are both accepted to ICML 2026 — lossy KV-cache compression for lossless inference, and sketch-and-walk sparse attention.
Nov 01, 2025 Word Salad Chopper appears at EMNLP 2025 — reasoning models waste a surprising share of their decoding budget on repetitions they already “know” are useless.
Jul 01, 2024 Two papers at ICML 2024Knowledge Graphs Can be Learned with Just Intersection Features and GNNs Also Deserve Editing.