High Performance Engineer (up to £140k)

Dex; Cambridge; On-site; GBP 120,000 - 180,000; Full time

Job summary

Dex in Cambridge is seeking a senior individual contributor to own performance across the full simulation and edge runtime stack. You will sit with founders who still write engine code, unblocking a talented early-career team while pushing code running on real robots.

You will profile, diagnose, and optimize high-throughput GPU physics simulation and edge pipelines, refactor critical paths, and mentor engineers through design reviews and paired work.

Qualifications

• Production-grade C++ performance optimization
• Experience in high-throughput/low-latency domains
• Mentoring engineers through pairing and code reviews
• Ability to work on-site with robotics hardware in Cambridge

Responsibilities

• Profile, diagnose, and optimize high-throughput GPU physics simulation, policy learning pipelines, and edge execution runtimes.
• Refactor critical paths to maximize hardware utilization, addressing memory layout, cache misses, SIMD utilization, and multi-threaded synchronization.
• Deploy and validate real-time execution pipelines directly on physical humanoid and quadruped robots in the lab.
• Set the standard for code quality and systems engineering by mentoring early-career engineers through pairing, design reviews, and clear task decomposition.
• Collaborate with founders to debate architecture, test hypotheses, and push raw simulation performance.

Skills

Modern C++

Hardware-aware parallelism

Profiling & memory layout

On-site collaboration

Job description

The role

Robots train inside simulation, and execution speed is the entire product. Founded by veterans who spent over a decade building industry-standard GPU physics engines at the world's leading hardware firm, this team is building the end-to-end physics, training, and real-time edge runtime stack for the next generation of physical robotics. They are backed by top-tier venture capital with a multi-year runway, working as a tight-knit engineering team in Cambridge alongside quadruped and humanoid hardware.

You will own performance across the entire simulation and edge runtime stack, finding where cycles disappear and taking them back. This is an elite senior individual contributor role, not a middle-management seat. You will sit alongside founders who still write engine code every day, serving as the senior technical anchor in the room to unblock and mentor a talented early-career team while pushing code that runs on real robots.

The work

• Profile, diagnose, and optimize high-throughput GPU physics simulation, policy learning pipelines, and edge execution runtimes.


• Refactor critical paths to maximize hardware utilization, tackling memory layout, cache misses, SIMD utilization, and multi-threaded synchronization.


• Deploy and validate real-time execution pipelines directly on physical humanoid and quadruped robots operating in the lab.


• Set the standard for code quality and systems engineering by mentoring early-career engineers through pairing, design reviews, and clear task decomposition.


• Collaborate directly with hands-on technical founders to debate architecture, test hypotheses, and push raw simulation performance.

What You Bring

• Deep mastery of modern C++ with proven experience making production systems measurably faster through hardware-aware parallelism, profiling, and cache-conscious memory layout.


• A background in domains where throughput or latency is the product itself, such as physics simulation, game engines, HPC, low-latency trading, or real-time embedded systems.


• Demonstrated experience lifting the technical bar of an engineering team through direct pairing, actionable code reviews, and structured task breakdown.


• Ability to work fully on site in Cambridge, UK alongside the founders, team, and physical robot hardware.