GPU Performance tooling engineer

Position Overview

• Location Type: Remote
• Job Type: Full-time
• Salary: Not specified
• Description: We are developing software to enhance the Deep Learning ecosystem and assist hardware engineers in building advanced Deep Learning parallel systems. We are seeking a candidate with a strong background in systems software development and experience with GPU performance tools and tracing frameworks like Perfetto. This role involves designing and implementing features for the Perfetto framework to simplify performance measurement for Rivos' internal and external users. The position may also involve in-depth work on the libraries that facilitate performance data collection. You will gain technical and organizational skills from industry experts, including best practices for writing performant and readable code, structuring and communicating projects, ideas, and progress, and collaborating effectively with the Open Source community. We are strong advocates for Open Source and Free software, contributing our improvements back to the projects we utilize.

Requirements

• Experience with Perfetto profiling framework: Required
• Ability to write code in C or C++: Required
• Understanding of computer architecture: Required
• Strong problem-solving skills: Required
• Excellent skills in problem solving, written and verbal communication: Required
• Strong organization skills, and highly self-motivated: Required
• Ability to work well in a team and be productive under aggressive schedules: Required

Optional Requirements

• Experience with profiling and optimizing low-level performance (memory bandwidth, latency, throughput) on GPU architectures.
• Familiarity with deep learning frameworks (TensorFlow, PyTorch, etc.).
• Ability to write code in Rust.

Responsibilities

• Develop and modify the Open Source Perfetto framework to enable software developers to improve the performance of their code.
• Work on underlying libraries and drivers to facilitate performance data collection.
• Ensure that performance monitoring overhead is minimized.
• Build tooling to facilitate measuring performance in various scenarios (on simulators, FPGAs, or real hardware).
• Write unit tests and benchmark tools to validate the performance and correctness of your changes.
• Stay current with advancements in the field.

Education and Experience

• Degree: Bachelor’s, Master’s, or PhD
• Field: Computer Engineering, Software Engineering, or Computer Science.

Application Instructions

• Not specified.