About the Role

The Performance Engineer will focus on optimizing managed language runtimes for ARM64 mobile-class CPUs, working across the entire runtime stack. This role involves identifying performance bottlenecks, improving compiler efficiency, and collaborating with CPU architects to enhance runtime workloads.

What You'll Do

Finding and ranking performance bottlenecks in interpreter, JIT/AOT, GC and runtime library code using sampling profilers, PMU counters, top-down analysis and flame graphs.
Improving interpreter and compiler tiers, including dispatch, inline caches, hidden classes, tier-up and deoptimisation policy, instruction selection, register allocation, inlining heuristics and check elimination.
Reducing startup time through snapshots, AOT images, lazy compilation and warm-up tuning.
Tuning allocation and collection for throughput, pause time and footprint, including barriers, generational and concurrent collectors, object layout and pointer compression.
Explaining runtime behaviour in hardware terms, such as indirect-branch misprediction, I-cache and iTLB pressure and load-to-use latency, and turning findings into software changes and hardware feedback.
Building benchmarking and regression-tracking pipelines with noise control and statistically sound A/B comparison.
Building and using LLM-agent pipelines to analyse profiles, propose patches and validate correctness.
Working with CPU architects on which hardware or ISA features would help runtime workloads.
Landing upstream patches, writing technical reports and contributing to patents and publications where appropriate.

Who You Are

Degree in Computer Science, Computer Engineering, Electrical Engineering or a related field, or equivalent practical experience.
Strong C and C++ (or Rust), with the ability to work productively in a large, performance-critical codebase.
Hands-on performance work in a language runtime, VM or compiler backend, with measured speedups you can walk through in detail.
Fluency with profiling tools such as perf, PMU counters and flame graphs, and the ability to read disassembly and relate it to source.
Solid grasp of a modern ISA (ARM64 or x86-64), calling conventions, the memory model and the cost of branches and cache misses.
Measurement rigour: you design experiments, control for noise and report results with appropriate statistics.

Bonus Points

Working knowledge of CPU microarchitecture, including out-of-order pipelines, branch prediction, caches, TLBs and prefetching.
ARM64 experience, particularly on mobile or embedded SoCs.
Contributions to an open-source runtime, compiler or GC.
Experience with LLM agents or ML-assisted tooling for code optimisation.
Familiarity with simulators such as gem5.
Publications or patents in runtimes, compilers or computer architecture.

Compensation

This is an inside IR35/PAYE contract, initially for 12 months and will require 100% onsite presence in Cambridge, UK.

Benefits

In accordance with local employment laws, applicants must have current, valid authorisation to work in the UK at the time of application.

Requirements

C and C++ proficiency

Strong skills in C and C++ (or Rust) are essential for working in a performance-critical codebase.

Performance optimization experience

Hands-on experience in optimizing language runtimes, VMs, or compiler backends is required.

Profiling tools expertise

Fluency with profiling tools such as perf and PMU counters is necessary.

ISA knowledge

A solid understanding of modern ISAs like ARM64 or x86-64 is important.

Nice to Have

CPU microarchitecture knowledge

Familiarity with CPU microarchitecture concepts such as out-of-order pipelines and branch prediction is a plus.

Open-source contributions

Experience contributing to open-source runtimes or compilers is desirable.

Benefits

Work authorization

Applicants must have valid authorization to work in the UK.