Python Software Engineer – MLOps Platform

About the role

Revolut is looking for a Python Software Engineer to join its AI team and build the core infrastructure that powers its MLOps platform. You will design and maintain scalable backend services that support the full AI lifecycle, from model training to inference, on GPU‑enabled environments.

Key responsibilities

• Design, develop, and operate scalable backend services and tooling for AI workloads.
• Profile and optimise model training and inference using PyTorch and CUDA‑enabled libraries.
• Improve GPU utilisation, memory consumption, and bandwidth across multi‑tenant, multi‑node clusters.
• Implement mixed‑precision execution, quantisation, efficient batching, gradient accumulation, and memory‑efficient attention.
• Create repeatable benchmarking practices, automated regression tests, and guide engineering teams on execution frameworks and GPU selection.

Required profile

• Proven experience designing and operating scalable backend systems in production (Linux, Kubernetes, containerised workloads) using Python.
• Hands‑on experience running, profiling, and optimising machine‑learning workloads built with PyTorch on NVIDIA GPUs.
• Strong understanding of GPU execution concepts, memory hierarchy, CPU/GPU synchronisation, and CUDA stream behaviour.

Required skills

• Python
• PyTorch
• CUDA
• NVIDIA GPUs
• Linux
• Kubernetes
• Containerisation (Docker)
• NCCL and multi‑GPU communication
• Mixed‑precision and quantisation techniques
• Benchmarking and performance profiling