Tenstorrent

Principal Engineer, Compilers

Full-Time in Toronto, CA

Tenstorrent is helping enable a new era in artificial intelligence (AI) and deep learning with its breakthrough processor architecture and software. The company’s mission is to deliver orders of magnitude better performance and efficiency for AI workloads from the datacenter to edge of Cloud by co-designing hardware, software and AI algorithms with our unique technology.

Tenstorrent’s architecture scales from datacenter servers to IoT devices with dramatically improved efficiency, flexibility, programmability compared to legacy accelerators including CPUs, GPUs, FPGAs, and TPU-type processors. It is developed by our world-class team with deep expertise in computer architecture, hardware design and verification, systems engineering, compilers, software development, and machine learning algorithms.

Our engineering-based culture is focused on achieving the highest levels of AI innovation across all of Tenstorrent’s technical disciplines. We constantly strive to blend best-in-class aspects of integrity, openness, diversity and collaboration throughout the company: from the CEO to the engineering leadership and to the newest employee who may be a recent college graduate. By joining Tenstorrent, you will be an integral part of a highly accomplished and distinguished team that has many years of experience at companies that include AMD, Arm, Intel and NVIDIA, and that thrives on delivering new, innovative products.

Based in Toronto with operations in Austin and California, Tenstorrent is growing quickly. And, we are proudly backed by top-tier Venture Capital firms including Real Ventures and Eclipse Venture Capital, as well as prominent industry luminaries.

Tenstorrent offers a highly competitive compensation package and benefits, and we are an equal opportunity employer.

Responsibilities:

  • In partnership with the leadership team, create and define the product roadmap and co-design of Tenstorrent's hardware and software stack
  • Develop machine learning graph compiler
  • Be the subject matter expert across one or more areas: scheduling, parallelization, memory allocation, data flow optimizations, optimizing kernels
  • Define the benchmark, analyze, and optimize performance of key machine learning applications across Tenstorrent's hardware and software stack
  • Develop performance analysis and estimation infrastructure that feeds into Tenstorrent compiler
  • Integrate the Tenstorrent software into leading machine learning frameworks
  • Provide technical guidance and direction to junior staff member
  • Work closely across teams to discover the hardware and software requirements of current and future machine learning applications

Qualifications & experience:

  • Deep understanding of IR and machine level compiler optimization techniques.
  • Extensive working experience with one or more: scheduling, parallelization, memory allocation, data flow optimizations, optimizing kernels
  • 7+ years of experience working with algorithms, data structures, and software development
  • C++, Python
  • Bachelors in Computer Science or Electrical/Computer Engineering or Engineering Science
  • Excellent verbal and written communication skills
  • Ability to work across multiple teams

Asset:

  • 2+ years of experience as a technical lead, developing and mentoring a team
  • Experience contributing to open-source projects, and demonstrated influence in the open community.
  • Masters/PhD in Computer Science or Electrical/Computer Engineering or Engineering Science
  • Software development in Python