ML / AI Platform Engineer

Dex; Greater London; On-site; GBP 270,000 - 330,000; Full time

Job summary

Dex is hiring an ML/AI Platform Engineer to design and operate scalable training and inference infrastructure in London / Europe. You will own the training stack, GPU clusters, and the platform tooling that drives the ML org.

The role emphasizes reliability, observability, and performance, with compensation up to £300k+ total compensation plus equity. You’ll work across frontier fintech and enterprise AI teams to optimize throughput and cost.

Qualifications

• Strong ML infrastructure or platform engineering background.
• Experience with large-scale training and serving systems.
• Focus on reliability, observability and performance.

Responsibilities

• Own training stack and GPU clusters for scale.
• Develop tooling for ML platform and production operations.
• Improve throughput, latency, and reliability of deployments.
• Take systems from initial build-out to steady production.

Skills

ML infra

Platform engineering

Distributed training

GPU optimisation

Observability

Tools

Kubernetes

CUDA

Linux

Perf profiling

Job description

ML / AI Platform Engineer (Train & Serve at Scale | up to £300k+ TC) · London / Europe

Location: London / Europe

Salary: Total compensation up to £300,000+ plus equity

Platform work gets treated as plumbing at a lot of companies. Not at the ones we work with - frontier labs, fintechs and enterprise-AI teams whose products live or die on training throughput, inference cost and cluster reliability. There, the platform engineer is one of the most consequential hires they make, and the comp reflects it. We know these teams day to day, so we can tell you where the infrastructure genuinely is the product.

The opportunities

Salaries across these roles run up to £300k+, with equity on top. You'd own the systems everything else stands on: the training stack, the GPU clusters, the inference layer, the tooling every ML engineer ships through. Impact is measurable in the units that matter - throughput, latency, reliability - and visible to the whole company when you move them.

You could work on

• Building and running training and inference infrastructure at scale


• Owning GPU and cluster performance, throughput and reliability


• Building the platform and tooling the ML org depends on


• Taking systems from first build-out to steady production operation

You may have

• A strong ML infrastructure or platform engineering background


• Distributed training, GPU optimisation or large-scale serving experience


• Comfort going low-level - kernels, profiling, parallelism - when it counts


• A production mindset: reliability and observability as first-class work