Acceleration Kernel Developer Lead
TenstorrentJob Title
Acceleration Kernel Developer Lead
Role Summary
Lead a team that develops and optimizes low-level compute and data-movement kernels for machine learning on Tenstorrent's networked RISC-V and AI processors. Drive performance tuning, profiling, and cross-team collaboration to deliver high-throughput, reliable software for ML workloads.
Experience Level
Senior. The role is a technical lead position; candidates will be assessed during interviews and offers aligned to assessed level.
Responsibilities
Primary responsibilities focus on kernel development, performance optimization, and team leadership.
- Lead and mentor engineers working on low-level, high-performance kernels for ML workloads.
- Design, implement, and optimize tensor compute and data-movement kernels in C/C++ for Tenstorrent architectures.
- Profile software and diagnose hardware–software bottlenecks using performance tools and runtime debugging.
- Collaborate with compiler, runtime, and ML framework teams to map model requirements to efficient low-level implementations.
- Define and enforce performance and quality standards for kernel code and releases.
- Contribute to hardware–software co-design discussions to improve execution efficiency on custom processors.
- Plan and prioritize work across the team to meet project milestones and product goals.
Requirements
Key qualifications and constraints for immediate effectiveness.
- Strong proficiency in C/C++ with experience writing high-performance, low-level code.
- Proven experience developing and optimizing tensor compute or data-movement kernels.
- Deep familiarity with performance profiling tools and debugging complex runtime environments.
- Solid understanding of machine learning frameworks and how models map to kernels and runtimes.
- Demonstrated technical leadership or team lead experience in a software or systems engineering context.
- Ability to translate ML requirements into performant low-level implementations and to drive cross-team coordination.
- Nice-to-have: GPU programming experience (CUDA, OpenCL), operating system internals, or prior work on compiler/runtime co-design.
- Employment contingent on eligibility to access U.S. export-controlled technology; candidates may need appropriate citizenship/permanent residency or prior license approval.
Education Requirements
Not specified.
About the Company
Company: Tenstorrent
Headquarters: Austin, Texas, United States
Tenstorrent is a technology company focused on designing innovative computing solutions. They are known for their expertise in the development of advanced hardware, including ASICs and SoCs, aimed at enhancing performance and efficiency in various applications.
