Master's Thesis: Advanced AI Accelerators

About the role

Huawei Sweden R&D, in the Baseband Lab in Stockholm, is offering a master's thesis project on advanced AI accelerators. The project focuses on the power and thermal analysis of emerging memory technologies in 3D-integrated AI accelerators, where 3D integration helps with the memory wall but can worsen thermal issues. The work starts in spring 2027.

Key responsibilities

• Develop a simulation framework for evaluating the power and thermal characteristics of 3D-integrated computing systems using conventional and emerging memory technologies such as advanced SRAM and MRAM.
• Integrate architectural simulators such as gem5 with power estimation models and thermal simulation tools such as HotSpot.
• Use the framework to study how memory technologies, reinforcement-based task scheduling and workload characteristics affect power consumption, performance and temperature distribution of the AI SoC.

Required profile

• Background in Computer Engineering or Electrical Engineering.
• Basic knowledge of computer architecture and programming.
• Strong spoken and written English.
• Able to work both in teams and independently, with attention to quality.

Required skills

• Desirable: familiarity with simulation tools (gem5, HotSpot, McPAT).
• Desirable: knowledge of memory technologies and 3D integration.