AI Performance Modeler Job in UK 2026 (£120,000 to £150,000 Annually) | Olix
Hey there! Olix is on the hunt for a world-class AI performance modeler to join their elite engineering team in London in 2026. We're offering a super competitive salary between £126,000 and £154,000 a year, plus equity and some fantastic benefits. This is a standout opportunity for international experts in computer architecture, performance modeling, and AI systems who want to make the UK their home base, with full visa sponsorship and complete relocation support to make the move smooth.
You'll be based right at Olix's headquarters in King's Cross, smack in the middle of Europe's hottest AI and deep-tech scene. Here, you'll help shape the future of optical and accelerator-based computing by diving deep into data-driven performance analysis.
About the Role
Picture this: You'll be in charge of creating and refining the analytical and simulation models that steer Olix's optical compute (OTPU) architecture and software plans. You'll build functional simulators and super accurate, cycle-by-cycle models to explore design options, spot bottlenecks, and guide key decisions on hardware and software.
This spot sits right where hardware architecture, software tools, and machine learning workloads meet—perfect for engineers who love crunching numbers, thinking big-picture, and iterating fast. Your insights will shape architectural choices, scheduling tactics, and the long-term direction of the platform.
About the Hiring Firm
Olix is a cutting-edge AI hardware company crafting groundbreaking optical computing systems to boost performance and efficiency on a massive scale. Based in King's Cross, London, they provide a top-notch work environment, the best equipment, and a culture where everyone feels like an owner through stock options.
Olix is all about welcoming talent from around the globe, with full UK visa sponsorship and relocation help to ensure a hassle-free transition. The company stands out for its bold mission, top-tier engineering vibe, and big investments in people, tools, and future-focused innovation.
Responsibilities
- Design, build, and keep up functional simulators for the OPTU subsystem and the whole compute pipeline
- Create and manage architectural and cycle-accurate performance models to check out throughput, latency, and utilization
- Dive into real workloads like LLMs, diffusion models, and graph tasks to gather detailed traces
- Run big design-space explorations to weigh trade-offs and steer hardware, software, and optical roadmaps
- Spot performance bottlenecks and suggest fixes through architecture or scheduling tweaks
- Develop Python and C++ tools for parsing traces, statistical analysis, and visualization
- Plug performance models into CI pipelines for automated performance checks
- Turn complex simulation results into clear, actionable insights for the team
Requirements
- At least 3 years of experience crafting performance or power models for CPUs, GPUs, ASICs, or accelerators
- Solid coding skills in C++ and Python, with hands-on experience in cycle-accurate or discrete-event simulators
- A good grasp of computer architecture, memory systems, interconnects, and performance analysis basics
- Familiarity with machine learning workloads and tools like PyTorch, TensorFlow, or JAX
- Ability to read RTL or hardware schematics and chat about micro-architectural trade-offs with hardware folks
- Strong skills in data analysis and visualization, plus the knack for explaining complex findings simply
- A Bachelor's degree or higher in Computer Science, Electrical Engineering, Physics, Applied Mathematics, or a related field
English
Arabic
French
Spanish
Portuguese
Deutsch
Turkish
Dutch
Russian
Romaian
Portuguese (Brazil)
Greek
