Lattice AI Research

Research focus

Distributed Systems & Hardware (2)Code Generation & Program Synthesis (1)Inference & Quantization (1)Architecture Design (Transformers, SSMs, MoE) (1)

Frequent co-authors

Jisheng Zhao (1)Shinnung Jeong (1)Chihyo Ahn (1)Hyesoon Kim (1)

Papers (2)

Apr 30, 2026

Jisheng Zhao +43w ago

CuLifter: Lifting GPU Binaries to Typed IR

Recovering type information from untyped GPU register files is the key to enabling effective binary analysis, unlocking reverse engineering and security analysis of proprietary GPU code.

Jisheng Zhao, Huanzhi Pu, Shinnung Jeong +2

Code Generation & Program Synthesis Distributed Systems & Hardware Inference & Quantization

Apr 20, 2026

Jie Liu +2Apr 20, 2026

AsyncSparse: Accelerating Sparse Matrix-Matrix Multiplication on Asynchronous GPU Architectures

Asynchronous GPU features like NVIDIA's TMA can unlock up to 6x speedups in sparse matrix multiplication, but only with careful kernel co-design.

Jie Liu, Huanzhi Pu, Zhiru Zhang

Architecture Design (Transformers, SSMs, MoE)Distributed Systems & Hardware Training Efficiency & Optimization

Search

Huanzhi Pu

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)