Huzama Ahmad

Ph.D. Candidate, KAIST AI.

I work on efficient language modeling — the architectures and systems that make large models cheaper to run at long context. Advised by Se-Young Yun in the OSI Lab. Recent work spans speculative decoding, sparse attention, and letting models control their own attention span.

Selected Publications

All

Venue: Preprint Year: 2026 Selected

Large Language Models Can Control Their Own Attention Span

Authors: Namgyu Ho^{* (equal contribution)}, Huzama Ahmad^{* (equal contribution)}, Woosung Koh^{* (equal contribution)}, Cicero Nogueira dos Santos, Tal Schuster, Se-Young Yun
Venue: Preprint Year: 2026 Selected

SpotAttention: Plug-In Block-Sparse Routing for Pretrained Long-Context Transformers

Authors: Huzama Ahmad, Se-Young Yun

Currently working on

All

Content-Aware Sparsity
Status: Active

Dates: May 2026 – Present

Stealth — details after publication

Huzama Ahmad

Selected Publications

Large Language Models Can Control Their Own Attention Span

SpotAttention: Plug-In Block-Sparse Routing for Pretrained Long-Context Transformers

Currently working on

Content-Aware Sparsity