Deep Learning Framework and Compiler Engineer

QUALCOMM, Inc.; Bengaluru; On-site; INR 2,500,000 - 6,000,000; Full time

Job summary

QUALCOMM, Inc. in Bengaluru seeks an experienced engineer to design machine frameworks for Adreno GPU accelerators used in computer vision and generative AI workloads.

You will optimize OpenCL/CUDA kernels, work on compiler graph-level optimizations, and integrate vendor SDKs such as CUDNN, TensorRT, and Adreno OpenCLML into inference frameworks. Strong C/C++, Python skills and ML framework experience are required.

Qualifications

• Bachelor's degree in Engineering, Information Systems, Computer Science, or related field and 4+ years of Systems Engineering or related work experience.
• Master's degree in Engineering, Information Systems, Computer Science, or related field and 3+ years of Systems Engineering or related work experience.
• PhD in Engineering, Information Systems, Computer Science, or related field and 2+ years of Systems Engineering or related work experience.

Responsibilities

• Design machine frameworks for Adreno GPU accelerators for computer vision and generative AI needs.
• Understand network architectures and how a deep learning framework optimizes it to efficiently run on target accelerator.
• OpenCL (or CUDA) kernel programming knowledge and optimization techniques to get best performance from hardware.
• Compiler optimization knowledge at graph level as well tensor IR; integrate vendor SDKs like CUDNN, Adreno OpenCLML, TensorRT in inference frameworks.

Skills

C/C++

Python

OpenCL/CUDA

GPU/Accelerator programming

DL frameworks

Parallel programming

Compiler optimization

Tensor IR

Education

Bachelor's degree in Engineering/CS/IS

Master's degree in Engineering/CS/IS

PhD in Engineering/CS/IS

Tools

TVM

MLIR

LLVM

XLA

Job description

General Summary:

• Design machine frameworks for Adreno GPU accelerator for computer vision and generative Ai needs
• Good understanding of network architectures, how a deep learning framework optimizes it to efficiently run on target accelerator
• OpenCL (or similar like CUDA) kernel programming knowledge and optimization techniques to get best performance from hardware
• Compiler optimization knowledge at graph level as well tensor IR
• Working experience with any one open source compiler frameworks like TVM, MLIR, LLVM, XLA etc
• Experience of understanding and integrating vendor acceleration SDK like CUDNN, Adreno OpenCLML, TensorRT, ArmNN inot inferenceing frameworks is a plus
• C, C++: Strong programming capability using advanced techniques to design and develop AI compilers and backends.
• Scripting: Strong expertise in Python with design, develop, release and maintain projects.
• AI Frameworks: Familiarity with other AI frameworks like PyTorch, TensorFlow, Hugging Face, etc.Machine Learning Knowledge:
• Understanding of machine learning principles and algorithms starting Computer vision to large language models and continuously update to new trends.Expertise to deep learning accelerator programming (GPU, NPU).
• Any parallel programming experience (Like CUDA, OpenCL, MKLDNN ..etc) is a plus.Experience with deep leaning compilers like Glow, TVM etc is a plus.

Minimum Qualifications:

• Bachelor's degree in Engineering, Information Systems, Computer Science, or related field and 4+ years of Systems Engineering or related work experience
• ORMaster's degree in Engineering, Information Systems, Computer Science, or related field and 3+ years of Systems Engineering or related work experience
• ORPhD in Engineering, Information Systems, Computer Science, or related field and 2+ years of Systems Engineering or related work experience