Responsibilities:
Research and develop key technologies of AI compilers and automatic
optimization tools
Research and develop software and hardware co-design methods and related
tools
For the IPU chip architecture, develop new AI compilation optimization
methods
Participate in the work of the AI compiler open source community
Requirements:
Strong background in computer architecture, in-depth understanding of at
least one AI chip architecture
Experience with assembly-level performance tuning on AI chips (including
GPUs or other AI chips)
Solid C++ system development ability
Understand the basic principles of deep learning training/inference
framework
Experience with AI processor adapting mainstream deep learning framework
is preferred
Bachelor degree or above in computer, electronics, communication or
mathematics
Excellent teamwork ability, self-motivation and results-driven
While we do our best to provide an accurate job description, as Graphcore's
business is dynamic the specifics of your job may change. Additional or
modified duties may be reasonably assigned to you based upon evolving business
needs
