Chenyang Ai

I am currently a PhD student at the Institute for Computing Systems Architecture (ICSA), School of Informatics, University of Edinburgh. My research centers on hardware microarchitecture and combines software scheduling, architecture exploration, and approximate computing to co-optimize systems for LLMs, CNNs, and other tensor-operator workloads.

PhD Student ICSA, School of Informatics University of Edinburgh chenyang_ai@stu.pku.edu.cn

Research Focus

Hardware Microarchitecture

Architecture design centered on efficient execution of tensor-heavy workloads across modern accelerators.

Software Scheduling

Cross-layer scheduling strategies that map computation, dataflow, and resources onto practical systems.

Architecture Exploration

Design-space studies that compare implementation tradeoffs across different operators, models, and execution styles.

Approximate Computing

Accuracy-efficiency tradeoff analysis for systems that target LLMs, CNNs, and other tensor-operator domains.

Current Position

My current work is anchored in hardware microarchitecture and extends upward into software scheduling and system-level optimization. The goal is to build coherent hardware/software co-design methods rather than treating architecture, compilation, and approximation as separate problems.

  • PhD student at the Institute for Computing Systems Architecture (ICSA).
  • Working within the School of Informatics at the University of Edinburgh.
  • Targeting LLMs, CNNs, and other workloads built around tensor operators.
  • Interested in balancing efficiency, programmability, and model quality.

CURRENT THEMES

Core directions

  • Microarchitecture-driven accelerator design.
  • Scheduling-aware hardware/software co-optimization.
  • Architecture exploration across different model and operator patterns.
  • Approximate computing for efficient tensor processing.
  • Research that connects theory, systems, and implementation tradeoffs.
Full profile