Role Overview
Join Arm's Manchester team building the compute platform for AI at scale. You'll work across the full stack — applications, frameworks, compilers, libraries, and assembly — to optimize agentic AI workloads on Arm Neoverse and next-gen cloud-server technologies like AGI-CPU. Collaborate with open-source communities and partners to accelerate AI performance.
Responsibilities
- Develop and optimize software for AI agentic compute platforms
- Debug complex issues across unfamiliar stacks and identify performance opportunities
- Work across the full stack: applications, frameworks, compilers, libraries, assembly
- Collaborate with cross-functional teams and open-source communities
- Own problems end-to-end without silos
Requirements
- Strong software engineering skills, preferably C++ and Python
- Eager to dive into unfamiliar codebases, debug complex issues, and find performance gains
- Familiarity with Linux development, profiling, debugging, and performance analysis
- Degree in computational/numerate subject or equivalent experience
- Right to work in the UK without sponsorship
Nice to Have
- ML frameworks/LLM inference: PyTorch, vLLM, SGLang, llama.cpp
- Heterogeneous CPU+accelerator systems
- Agentic AI frameworks and protocols
- End-to-end system performance analysis
- CPU bottlenecks in agentic workloads
- Databases, vector search, retrieval systems
- Containers and cloud/server infrastructure
- Parallel programming, collective communications, distributed AI
- Open-source community engagement
- SIMD technologies: Neon, SVE, SME
Benefits
- Attractive reward package with generous stock awards
- Progressive leave: parental leave, care days, 4-week sabbatical every 4 years
- Flexible hours and hybrid working
- Open, inclusive culture supporting professional and personal growth