Software Engineer, Tenstorrent
Metal Runtime, API & Abstractions
$100-500k
OTE
Next-generation computing company building computers for AI
Be an early applicant
Next-generation computing company building computers for AI
1001+ employees
Be an early applicant
$100-500k
OTE
1001+ employees
Company mission
To bring unique solutions to the problems facing AI and machine learning.
Role
Who you are
- If you enjoy pushing performance, working close to the metal, and designing abstractions that make complex systems usable without sacrificing control, this is your kind of role
- We welcome candidates at various experience levels for this role. During the interview process, candidates will be assessed for the appropriate level, and offers will align with that level, which may differ from the one in this posting
- Strong coder in C or C++, and enjoy working close to hardware
- Comfortable thinking in terms of threads, processors, and memory movement
- Care deeply about API design, abstraction boundaries, and developer experience in low-level systems
- Experienced (or interested in) designing libraries or interfaces used by other engineers
- Think about long-term maintainability, usability, and clarity when designing systems
- Curious about how runtimes and bare-metal software come together to power AI
- Collaborative by nature and enjoy solving tough problems with smart people
Desirable
- Experience designing or evolving non-trivial C++ APIs used by other engineers
- Familiarity with APIs for accelerator systems (e.g., CUDA, SYCL, Vulkan, or similar)
What the job involves
- As a Software Engineer on the Metal Runtime team at Tenstorrent, you’ll work on the low-level software that powers our AI accelerators. You’ll design fast, efficient runtime systems that run close to the hardware, and define the host and device APIs that expose these capabilities to the rest of the software stack
- We believe APIs are a core part of systems design: they encode hardware semantics, performance tradeoffs, and concurrency models, and they live longer than any single implementation
- Design and evolve host and device APIs that expose Metal runtime capabilities clearly and efficiently
- Translate low-level hardware behavior into intuitive, high-performance abstractions
- Collaborate closely with runtime, kernel, and higher-level software teams to align API design with real-world usage
- Ensure APIs balance performance, control, safety, and ease of use
- Contribute to runtime implementation where needed to support API design and integration
- Debug, analyze, and improve usability and performance across the software stack
- How to design APIs and abstractions for cutting-edge AI hardware
- How to balance performance, control, and usability in low-level system design
- How to translate hardware capabilities into clean, developer-friendly interfaces
- Best practices for building maintainable, high-performance systems software
- How software and hardware co-evolve in a fast-paced AI startup
Our take
Tenstorrent is a US- and Canada-based semiconductor startup specializing in processors for AI workloads. Rather than building gneral-purpose CPUs or GPus, the company develops custom chips and sofrware optimized for running moder AI models. As demand for AI accelerates, and as a global chip shortage reshapes the semiconductor industry, these trends have become strong tailwinds for Tenstorrent.
The company’s unique approach is to build for this future software, which it believes will be written at a higher level. Combining this with conditional computing, Tenstorrent supports a scalable distributed network approach that is analogous to a human brain.
While the market for specialized AI hardware is destined to be gigantic, the company has to compete with a plethora of hopeful startups (as well as major tech companies like NVIDIA and Qualcomm). That said, Tenstorrent boasts an excellent trajectory so far, with its latest moves including unveiling its high-end Wormhole AI processors, and a partnership with LG which will see the development of chips for "Affectionate Intelligence".
Kirsty
Company Specialist at Welcome to the Jungle
Insights
52% employee growth in 12 months
Company
Funding (last 2 of 7 rounds)
Dec 2024
$693m
SERIES D
Jun 2024
$300m
LATE VC
Total funding: $1.3bn
Company HQ
Flemingdon Park, Toronto, ON
Leadership
Jim Keller
(CEO)Previously Senior VP of Silicon Engineering at Intel Corporation, VP of Autopilot Hardware Engineering at Tesla, Corporate VP and Chief Architect at AMD, and VP of Engineering at Apple.
Salary benchmarks
We don't have enough data yet to provide salary benchmarks for this role.
Submit your salary to help other candidates with crowdsourced salary estimates.
Share this job
View 19 more jobs at Tenstorrent