Kernel Engineer - High-Performance ML/HPC on Custom AI Chip

Cerebras; United States; On-site; USD 100,000 - 130,000; Full time

Benefits offered by this job

Opportunity to publish open-source AI research

Work with one of the fastest AI supercomputers

Non-corporate work culture

Job summary

Cerebras is looking for a Kernel Engineer to innovate cutting-edge AI and HPC workloads. You'll develop high-performance software solutions optimizing deep learning operations leveraging our custom processor architecture.

Your role will involve building a library of algorithms and collaborating with top talents on the design and validation of ML and HPC kernels. Join Cerebras to be part of an exciting journey crafting breakthrough AI technologies.

Qualifications

• Comfortable learning new hardware architectures.
• Experience in kernel development and testing.
• Familiarity with parallel algorithms and distributed memory systems.

Responsibilities

• Develop specifications for ML and linear algebra kernels.
• Debug kernel library using assembly and CSL routines.
• Measure software performance to inform design decisions.
• Integrate unit and system testing methodologies.
• Evolve kernel library architecture for computational challenges.

Skills

C++ programming

Python programming

Debugging skills

Library and API development

Understanding of hardware architecture

Education

Bachelor’s, Master’s, PhD in Computer Science or related fields

Tools

TensorFlow

PyTorch

Job description

Cerebras is looking for a Kernel Engineer to innovate cutting-edge AI and HPC workloads. You'll develop high-performance software solutions optimizing deep learning operations leveraging our custom processor architecture.

Your role will involve building a library of algorithms and collaborating with top talents on the design and validation of ML and HPC kernels. Join Cerebras to be part of an exciting journey crafting breakthrough AI technologies.