Lead AI Infrastructure & Distributed Systems Engineer

LinuxRecruit; Greater London; On-site; GBP 100,000 - 140,000; Full time

Job summary

LinuxRecruit is building a foundational AI system for physical intelligence, aiming to train large models efficiently on GPU clusters from a centrally located London base. The role demands hands-on coding, system design, and full ownership of the training infrastructure in an early-stage startup setting.

You will work at the intersection of HPC, GPU orchestration, and MLOps, crafting scalable systems, optimizing throughput, and driving architectural decisions that unlock rapid experimentation

Qualifications

• Strong production background with AWS, Kubernetes, Slurm, PyTorch.
• Deep hands-on experience with GPU compute optimisation, cluster scheduling, and high performance networking.
• Experience in MLOps, AI infrastructure, or HPC; robotics experience not required, though CV pipelines are advantageous.

Responsibilities

• Own architectural decisions for core model training infrastructure from day one.
• Design scalable distributed systems from scratch for large GPU compute clusters.
• Code daily and reduce training bottlenecks to accelerate research cycles.

Skills

GPU compute optimisation

Cluster scheduling

High performance networking

MLOps

Distributed training

Tools

AWS

Kubernetes

Slurm

PyTorch

Distributed training frameworks

Job description

Building foundation model style AI for physical intelligence and pioneering a system that allows robots to learn complex physical tasks from just a single human demonstration (think "prompting" a robot in the physical world).

Having recently closed a $40M Seed round backed by premier global investors, (marking one of the largest robotics seed investments in European history), they are intentionally keeping their team small, elite, and talent dense. Their primary bottleneck is no longer collecting data, it is the raw speed, scale, and efficiency of training their models.

This business is looking for an elite engineer to "run the show" for their core model training infrastructure. Sitting at the intersection of high performance computing, GPU orchestration, and MLOps, this individual will professionalise the company's infrastructure, eliminate training bottlenecks, and drastically accelerate research cycles.

This is a deeply technical, hands on position. Rather than managing people or executing traditional DevOps/SRE maintenance, the successful candidate will be coding daily, designing scalable distributed systems from scratch, and squeezing maximum throughput out of large scale GPU compute clusters.

The founders are seeking a high agency engineer who thrives in early stage startup environments and prefers broad systems ownership over narrow specialisation.

• Technical Expertise: Strong production background with AWS, Kubernetes, Slurm, PyTorch, and distributed training frameworks. Deep hands‑on experience with GPU compute optimisation, cluster scheduling, and high performance networking is essential.
• Relevant Background: Experience in MLOps, AI infrastructure, or HPC. Prior robotics experience is not required, though experience with computer vision pipelines is highly advantageous compared to text‑only LLM backgrounds.
• Mindset: A desire to join a founding level team in person in central London, take full ownership of architectural decisions from day one, and capture substantial upside through equity.