Papers
1
Total Citations
17
H-Index
1
About
Qi Guo is a leading figure in the field of efficient AI computing, with a primary focus on bridging the gap between massive-scale machine learning models and resource-constrained edge devices. His most impactful work centers on hardware-software co-design for deploying large language models (LLMs) on mobile and embedded platforms. In his seminal 2024 paper, "Cambricon-LLM: A Chiplet-Based Hybrid Architecture for On-Device Inference of 70B LLM," Guo introduced a novel chiplet-based architecture that enables the inference of 70-billion-parameter models on devices like smartphones and robotics—a feat previously thought impossible. This work, already garnering 17 citations in its first year, addresses the critical challenge of single-batch computing with extremely low arithmetic intensity, a key bottleneck for real-time AI applications. Guo’s contributions are not merely theoretical; they offer practical pathways to enhanced user privacy and network resilience by keeping powerful AI capabilities local. His research is pivotal for the next generation of intelligent, autonomous systems, making him a prominent voice in the intersection of computer architecture and applied machine learning.
Research Focus
Key Achievements
Top Papers
- 1