Performance Engineer

source.dev

London

On-site

NOK 1,142,000 - 1,649,000

Full time

7 days ago

Be an early applicant

Get more replies from employers

Send a job-specific resume in minutes.

Job summary

Source.dev is seeking a Performance / Systems Engineer to push the speed limits of Rust, Linux and I/O. This role focuses on dissecting execution paths, reducing kernel-space costs, and squeezing maximum throughput from SourceFS.

You will design robust benchmarks, optimize memory layouts, and work with a collaborative team that values End-to-End Ownership and practical problem solving. Join us in shaping high-performance device software tooling.

Qualifications

Deep understanding of computer architecture and memory systems.
Experience writing high-performance Rust or C/C++ code.
Interest in learning Rust if not already.
Experience with large-scale build systems and developer tooling.
Ownership mindset and collaborative, empathetic teamwork.

Responsibilities

Profile, trace, and eliminate latency across CPU, memory, and disk I/O bottlenecks in Rust systems.
Tune Linux kernel interactions, optimize FUSE drivers, VFS caches, io_uring.
Improve memory and cache efficiency; minimize copies across boundaries.
Design reproducible benchmarking to measure performance regression.
Maximize build performance on large graphs (Android, Yocto) on high-core hardware.

Skills

Rust

C/C++

Performance optimization

System programming

CPU architecture

Hyper-optimized code

Tools

perf

eBPF

valgrind

io_uring

Linux kernel

Job description

UK | Full-Time

The Mission: Source.dev

Our founders spent over a decade at Google shaping the Android operating systems - building and shipping the software that now runs on billions of devices. They didn't just use Android; they led teams responsible for the Android Virtualization Framework, Security, and the Android Runtime (ART).

Through that journey, they saw a quiet but massive crisis: as device software grew in complexity, the tools used to build it remained stuck in the past. Today, a standard Android checkout exceeds 200GB of code, and builds can take hours. This \"wait time\" is the single biggest tax on innovation.

Source.devwas founded to reimagine software development workflows for device makers. We’re building a new generation of software tools that make building complex device software feel as simple as building a webpage - starting with SourceFS.

The Core Product: SourceFS

SourceFS is a high-performance virtual filesystem designed to turn multi-hour workflows into minutes.

10x Faster Builds: Intelligent caching and virtualization that only materializes files when they are actually needed.

10x Lower Costs: Enabling a standard 16-vCPU machine to outperform even a 128-vCPU machine.

True Engineering Productivity: Replaying hundreds of thousands of build steps instantly when inputs match.

The Cultural Anchor: Own It

We operate with a high degree of trust and a shared value of End-to-End Ownership:

\"We own our work end-to-end, from spark to finish. It’s about more than just tasks; it’s the dedication to see things through, staying curious to solve problems, thinking from first principles and ensuring we fix our customer’s problems in the best possible way.\"

The Role: Pushing the Speed Limits of Rust, Linux & I/O

We are looking for a Performance / Systems Engineer to join us at Source.dev. In this role, your primary metric of success is speed. You will dissect execution paths, eliminate kernel-space context-switching costs, optimize memory layouts, and ensure SourceFS runs as close to the hardware limits as physically possible.

What You’ll Work On

System Bottleneck Hunting: Profile, trace, and eliminate latency across CPU, memory, and disk I/O bottlenecks in Rust systems code using tools like perf, eBPF, valgrind, and custom microbenchmarks.

Kernel & File System Tuning: Deep dive into Linux kernel interactions, optimizing FUSE drivers, VFS caches, asynchronous I/O (io_uring), memory-mapped execution, and thread scheduling.

Memory & Cache Efficiency: Optimize data structures for cache locality, minimize allocation overhead, and eliminate unnecessary memory copies across boundary transitions.

Benchmarking Infrastructure: Design robust, reproducible benchmarking suites to measure micro- and macro-performance regression in high-concurrency environments.

Scaling Build Primitives: Squeeze peak performance out of massive build graphs (Android, Yocto, and more) running on high-core hardware configurations.

Ideal Candidate Profile:

The Performance Systems Mindset: You have a deep understanding of computer architecture (CPU caches, memory barriers, SIMD, system calls) and proven experience writing hyper-optimized code in Rust or C/C++.(or other). No prior experience with Rust required, just a willingness to learn is all we are looking for.

Scale-Obsessed: You have experience with (or a burning desire to fix) large-scale build systems and developer tooling.

Impact-Driven: You embody the \"Own It\" mentality - you see a friction point in the developer journey and you don’t stop until it’s gone.

Curious & Collaborative: You enjoy translating real-world customer problems into robust, production-grade solutions.

Low Ego, High Impact: Beyond being a brilliant coder, you are a thoughtful and empathetic teammate - someone who leads with kindness

Feedback Culture Champion: You embrace feedback and provide it thoughtfully; you see open, honest communication as the foundation of great teamwork.

A Note on Our Search

We are looking for the best talent, wherever it may be. If you’re on the fence about whether you’re a perfect fit, please do apply - we value diverse backgrounds and unique problem-solving approaches.

Why Join Source.dev?

AI-Native Infrastructure: We merge deep systems expertise with AI to solve the \"impossible\" - using AI to automate complex merges and conflict resolution for the next generation of devices.