ML Research Engineer, Performance Optimization

About the Role

Join our Research & Engineering team as an ML Research Engineer focused on performance optimization, where you will architect and implement cutting-edge solutions to redefine AI efficiency on various hardware platforms. This remote, full-time role offers the opportunity to make a significant impact on the frontier of intelligent systems, working across a global team.

Responsibilities

• Design and implement high-performance, custom GPU kernels for ML training and inference.
• Utilize low-level profiling tools to tune and optimize kernel performance.
• Integrate optimized GPU kernels into ML frameworks like PyTorch, bridging high-level models with low-level hardware.
• Apply a deep understanding of memory hierarchy to optimize compute and memory-bound workloads.
• Develop fine-grain optimizations targeting specific hardware architectures, such as tensor cores.

Requirements

• Experience writing high-performance, custom GPU kernels for training or inference.
• Proficiency with low-level profiling tools and demonstrated ability to tune kernels effectively.
• Experience integrating GPU kernels into ML frameworks (e.g., PyTorch).
• Solid understanding of memory hierarchy and optimization techniques for compute and memory-bound workloads.
• Demonstrated ability to implement fine-grain optimizations for target hardware, including tensor cores.

Role Impact & Work Environment

As part of the Research & Engineering team, your work will directly contribute to building efficient AI systems that operate on-device, at the edge, and under real-time constraints, pushing the boundaries of intelligent system architecture. This is a full-time, remote position open to candidates in the United States, Austria, Canada, France, Germany, Netherlands, Switzerland, and the UK.