About Me

My research spans the intersections of deep learning, hardware architecture, and computer networks, with a particular focus on developing efficient algorithms and systems for deploying artificial intelligence. I have extensive experience in optimizing deep neural networks (DNNs) through various approaches, including systolic array implementations, sparse architectures, and quantization techniques. My work has contributed to both theoretical frameworks and practical implementations for accelerating DNN inference on edge devices, especially in distributed network environments. Recently, I have expanded my research to address efficiency challenges in emerging AI systems, particularly Large Language Models (LLMs) and multi-modal architectures, as evidenced by my work on speculative decoding and transformer optimization techniques.


News


Recent Publications

PipeSpec: Breaking Stage Dependencies in Hierarchical LLM Decoding

Bradley McDanel, Sai Qian Zhang, Yunhai Hu, Zining Liu
ACL Findings, 2025.
preprint
code

a diagram of a computer system
Speculative Decoding and Beyond: An In-Depth Review of Techniques
Yunhai Hu, Zining Liu, Zhenyuan Dong, Tianfan Peng, Bradley McDanel, and Sai Qian Zhang
Beyond Trusting Trust: Multi-Model Validation for Robust Code Generation

Bradley McDanel
UMBC CODEBOT ‘25 Workshop
paper
slides

AMUSD: Asynchronous Multi-Device Speculative Decoding for LLM Acceleration

Bradley McDanel
IEEE International Symposium on Circuits and Systems (ISCAS), 2025.
preprint
code

a series of diagrams showing different types of circuiting