At a Glance
- Tasks: Advance accelerator programming and optimise low-level kernels for cutting-edge AI architectures.
- Company: Join KRAI, a leader in AI benchmarking and optimisation based in Cambridge, UK.
- Benefits: Competitive salary, collaborative environment, and opportunities to work on groundbreaking projects.
- Other info: Dynamic team contributing to open-source and leading AI competitions.
- Why this job: Be at the forefront of AI technology and make a real impact in the industry.
- Qualifications: Advanced degree in Computer Engineering/Science and 3+ years of experience with accelerator hardware.
The predicted salary is between 70000 - 90000 £ per year.
We are looking for research-oriented engineers to advance the state-of-the-art of accelerator programming. In particular, we are aiming to supercharge the development of low-level kernels for a variety of accelerator architectures. True challenge comes from mapping the ever evolving workloads (e.g. AI models) onto the ever evolving hardware (e.g. AI accelerators) under the infinite space of quality/performance/cost/etc requirements and trade-offs. Our approach blends the timeless pursuit of Pareto optimality with systematic engineering. Our goal stretches beyond producing functionally correct solutions, but reaching for ones that surprise and delight even seasoned professionals (e.g. ourselves or our customers!).
Qualifications
- You should be intimately familiar with traditional tools of the trade such as compilers, debuggers, profilers, simulators, roofline models, etc.
- Not be afraid (or ashamed!) to use frontier AI techniques to help solve complex engineering challenges.
- An advanced degree in Computer Engineering/Science or Natural Sciences.
- At least 3 years of experience optimizing compute-intensive workloads on accelerator hardware (e.g. GPUs or FPGAs).
- Experience with optimizing inference across the full stack (“neuralware”/middleware/hardware).
- Strong communication, collaboration, and leadership skills.
Background
Headquartered in Cambridge, UK (“Silicon Fen”), KRAI is a purveyor of premium benchmarking and optimization solutions for AI Systems. We work with a variety of accelerator architectures, with some of them years before they reach maturity. We successfully combine our knack for performance engineering with rigorous methodology and artistic flair. We actively contribute to open-source and research efforts. We are proud to be a Founding Member of MLCommons, the parent organization of the industry-leading MLPerf competitions, sometimes called “The Olympics of AI Systems”. Our team has participated in over 20 MLPerf rounds since 2019, a record only matched by NVIDIA. As an “Olympic coach” to accelerator designers (e.g. Qualcomm, Google), server manufacturers (e.g. HPE, Dell, Lenovo), and cloud providers (e.g. Nebius), we have contributed to some of the fastest and most energy efficient results in the history of MLPerf.
The Art of Accelerator Programming employer: HiPEAC
KRAI is an exceptional employer located in the heart of Cambridge, UK, where innovation thrives in the vibrant tech hub known as Silicon Fen. We foster a collaborative and dynamic work culture that encourages creativity and professional growth, offering our engineers the opportunity to work on cutting-edge projects that push the boundaries of AI technology. With a commitment to open-source contributions and a track record of excellence in industry competitions, we provide a unique environment for those looking to make a meaningful impact in the field of accelerator programming.
