GPU HW Research Engineer

Qualcomm•San Diego, CA

About The Position

Qualcomm’s GPU Research Team is seeking innovative GPU architects to advance state-of-the-art capabilities in Artificial Intelligence (AI), Machine Learning (ML), and General-Purpose GPU (GPGPU) computing. This is a unique opportunity to design next-generation GPU architectures that power everything from mobile devices to Windows on Snapdragon (WoS) compute platforms and large-scale data center GPUs.

Requirements

Strong understanding of GPU architectures, programming models, and application domains.
Proficiency with APIs such as OpenCL, CUDA, Vulkan, or Direct3D 12.
Experience with hardware design flow and basic knowledge of Verilog (or VHDL).
Exposure to hardware simulation/emulation tools and waveform analysis.
Basic understanding of GPU memory and cache design hierarchy.
Solid programming skills in C/C++.

Nice To Haves

Familiarity with Python is a strong plus.
Hands-on experience in development, debugging, and optimization of CUDA and OpenCL kernels, as well as GPU compute shaders.
Deep understanding of quantization techniques and data types for large language models (LLMs) and large vision models (LVMs).
Experience designing GPU hardware features for ML/AI acceleration.
Experience with open-source projects such as llama.cpp, vLLM, or similar frameworks is a big plus.

Responsibilities

Collaborate with other GPU architects to design new hardware features and enhance existing GPU architectures for accelerating GPGPU, ML, and AI workloads.
Develop architectural solutions for diverse platforms, including mobile, WoS, and data center GPUs.
Work closely with software teams, hardware design teams, standardization bodies, and internal/external partners to deliver cutting-edge solutions.
Participate in open-source GPGPU/ML/AI projects and contribute to industry-leading initiatives.
Influence the evolution of GPU capabilities for advanced use cases such as large language models (LLMs) and large vision models(LVMs).