Alireza Khadem

AI Researcher

Sunnyvale, California, United States7 yrs 1 mo experience

Highly Stable

Key Highlights

Expert in hardware-software co-design for ML applications.
Proven track record in performance optimization for TPU datacenters.
Innovative solutions in accelerator design and memory systems.

Stackforce AI infers this person is a specialist in AI/ML and HPC with a focus on performance optimization and accelerator design.

Contact

Skills

Core Skills

Accelerator DesignPerformance OptimizationHardware-software Co-design

Other Skills

CC++CUDAMemory System OptimizationHPC ApplicationsML ModelsPerformance CharacterizationPerformance ModelingNeural Network Accelerator DesignFPGA ImplementationVerilogVHDLSystemVerilogUniversal Verification Methodology (UVM)

About

I am a Research Scientist on the GenAI Efficiency team at Google, where I work on improving the performance and efficiency of Gemini serving on Google TPU datacenters. I earned my PhD in Computer Science and Engineering from the University of Michigan, where my research focused on hardware–software co-design. My work centered on memory system optimization and the performance characterization, modeling, and optimization of HPC applications and ML models, including LLMs, CNNs, and GNNs. Previously, I interned at Apple and Microsoft Research, where I worked on design space exploration for LM serving on large-scale clusters and performance modeling of communication primitives.

Experience

Google

Research Scientist

Nov 2025 – Present · 5 mos · Sunnyvale, CA · On-site

I work in the GenAI Efficiency team to enhance the performance of Gemini serving on large-scale TPU datacenters.

Accelerator DesignCC++CUDAPerformance Optimization

Microsoft

Research Intern

Jun 2024 – Aug 2024 · 2 mos · Redmond, Washington, United States · On-site

Performance Modeling of Language Models

Performance ModelingPerformance Optimization

Apple

Design Verification Intern

May 2021 – Aug 2021 · 3 mos · Cupertino, California, United States

University of michigan

2 roles

Graduate Student Instructor

Aug 2020 – May 2021 · 9 mos · Ann Arbor, Michigan, United States

Graduate Student Instructor (GSI) of the Computer Architecture, Advance Topics in Computer Architecture, and Applied Parallel Programming with GPUs.

Graduate Student Research Assistant

May 2020 – Oct 2025 · 5 yrs 5 mos · Ann Arbor, Michigan, United States

My research focused on hardware-software co-design. On the hardware side, I optimized memory systems, proposed Processing-in-Memory solutions and designed domain-specific accelerators. From the software perspective, I characterized mobile applications, High Performance Computing simulations, and Machine Learning models such as Large Language Models (LLM), Convolutional, and Graph Neural Networks.

Hardware-Software Co-DesignMemory System OptimizationHPC ApplicationsML ModelsPerformance Optimization

University of tehran

Research Assistant

Sep 2017 – Dec 2018 · 1 yr 3 mos · Tehran, Iran

I proposed a power-efficient neural network accelerator to exploit weight redundancy in the neural network models. I also implemented this architecture on the FPGA SoC to detect a digit written on a sheet (MNIST dataset) through a camera connected to the FPGA. This idea is presented as a paper in IEEE Micro 2019.

Neural Network Accelerator DesignFPGA ImplementationAccelerator Design