DevOps Engineer — HPC & GPU Platform (Remote, Paris-based)

Overview

We are looking for a DevOps engineer with a strong software development background to join a distributed GPU compute platform project for a leading UK SaaS company (fintech/enterprise planning space).

The context

You will be embedded in a senior engineering team building a greenfield GPU-accelerated compute platform on AWS. The central SRE team manages the underlying infrastructure — your role is to build the tooling on top of it.

What You Will Work On

• Build GPU benchmarking frameworks on AWS: scheduling benchmark runs, collecting and storing results, enabling performance comparison across versions
• Develop correctness validation tooling: automated testing of numerical accuracy of GPU compute outputs against reference results
• Implement distributed observability across all platform services: structured logging, distributed tracing (Pulsar), performance metrics
• Contribute to broader HPC coding tasks alongside the engineering team

What we are looking for

• Strong Python or Go developer — you write real application code, not just scripts
• Experience with observability tooling (Prometheus, Grafana, distributed tracing)
• Comfortable with AWS (EC2, IAM, VPC) and CI/CD pipelines
• HPC or GPU environment experience is a strong plus — Slurm, compute clusters, GPU workloads
• ENSIMAG, Centrale, INSA, X or equivalent engineering background preferred
• English fluent — the team is distributed across France and the UK

Modalities

• 100% remote, 1 day/week in London
• Start: May 2026
• Contract: freelance/portage
• Rate: competitive, based on experience