F

Fractile

Specialized AI inference chips and software

Engineering Manager - Compilers

Engineering LeadershipConfirmed live in the last 24 hours
Full-timeIn-person

About the Job

Requirements

  • Proven experience managing a high-performance development team and delivering consistently outstanding outcomes.
  • Experience in compiler development; experience with machine learning compilers and technologies such as MLIR or TVM is advantageous.
  • Up-to-date knowledge of machine learning techniques and technologies.
  • A good understanding of computer architecture and performance optimisation techniques.
  • Excellent C and C++ skills and solid experience with industry-standard development tools and technologies.
  • Comfort making decisions with incomplete information and adapting to ambiguity.
  • Ability to understand the layers above and below the compiler stack and make sound decisions at system boundaries.

Responsibilities

  • Build and lead a team of 10+ engineers developing the compiler stack for Fractile's AI acceleration hardware.
  • Enable industry-leading large language model inference performance on Fractile's accelerators.
  • Shape the architecture of a compiler from first principles, including IR design, lowering strategy, and execution model.
  • Make architectural decisions as the target hardware progresses from specification to simulator to early silicon.
  • Work closely with hardware, driver, runtime, kernel, and machine learning engineers and researchers across the stack.
  • Hire and grow the compiler team, and deliver software for deployment in the company's systems.

Desired Qualifications

  • Experience with machine learning compilers and industry-standard technologies such as MLIR or TVM.

About the Company

Fractile

Specialized AI inference chips and software

Fractile designs specialised AI inference chips, systems, and software to run very large AI models. Its approach packs DRAM directly on top of logic to boost memory bandwidth and capacity, offering an alternative to HBM and SRAM-based designs. The company operates full-stack across research, hardware, software, and systems, delivering an integrated hardware-software stack. Its goal is to enable scalable, high-performance model inference by meeting demanding memory needs with a complete solution.

Company size
51-200
Company stage
Series C
Total funding
$984.1M
Headquarters
London, United Kingdom
Founded
2022

Recently Posted Jobs

Explore more opportunities at Fractile

See more jobs