A complete application in a minute
Tailored resume and cover letter, ready to send.
Callosum in Greater London is seeking a Member of Technical Staff for Systems Simulation & Performance Modelling to build simulators predicting AI hardware performance from single accelerator to multi-rack systems.
The ideal candidate has systems simulation or performance modelling background, deep knowledge of computer architecture or distributed systems, and hands-on profiling and tracing. Experience in C++ or Rust and a focus on modelling accuracy and trade-offs are valued.
The role is about building and extending simulators that predict how AI hardware performs from a single accelerator up to multi rack systems. A strong candidate brings systems simulation or performance modelling background, solid grasp of computer architecture or distributed systems, and hands on profiling and tracing work. Systems programming in C++ or Rust and clear thinking about modelling accuracy and trade offs matter more than any single tool.
How it fits the AI era
AI workloads now spread across many specialised chips, so predicting how they behave together before hardware exists becomes central. Simulation and modelling guide which hardware and execution choices get made, and that judgement stays human. The work pairs automated optimisation with careful measurement of real systems to keep models honest.
How it builds AI experience
The role offers direct exposure to heterogeneous accelerators from several vendors and to rack scale performance questions. Experience grows through building predictive models, validating them against traces from real deployments, and iterating on fidelity versus cost. That grounding in hardware aware performance work is useful for anyone moving toward AI infrastructure roles.
Worth showing
This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.
About Callosum
Callosum builds infrastructure that routes artificial intelligence workloads to the specialized model and silicon chip most suitable for each task, per the company.