Staff Engineer, AI Infra & Distributed Storage (Amsterdam Hybrid)

Together; Amsterdam; Hybrid; EUR 140,000 - 210,000; Full time

Job summary

Together AI is looking for a Staff Engineer to design, operate, and scale multi-petabyte storage systems and Kubernetes-native storage operators for AI workloads. You will work on high-throughput, low-latency data paths, multi-terabyte caching architectures, and integration of cutting-edge storage technologies for AI training and inference.

You will collaborate on securing production-grade storage with strict multi-tenancy, QoS, and automated provisioning while optimizing costs across thousands

Qualifications

• BS/MS in Computer Science, Engineering, or equivalent practical experience.
• 8+ years in storage engineering, multi-petabyte scale.
• Strong coding skills in Go and Python; production-grade tooling.

Responsibilities

• Architect and implement the storage roadmap for Together AI and scale GPU fleet.
• Integrate Vast, Weka, Ceph; optimize via automated tiering and lifecycle policies.
• Develop caching and tiered storage to achieve extreme IOPS and throughput.
• Tune L2/L3 network storage isolation for multi-tenancy in production.
• Code Kubernetes storage operators and controllers for provisioning and quotas.
• Engineer end-to-end data paths to 10+ GB/s per GPU node; multi-tier caching.
• Benchmark and profile to contribute to open-source storage projects.

Skills

Go

Python

Kubernetes

Linux

Distributed storage

Education

BS/MS in Computer Science or Engineering

Tools

Ceph

WekaFS

Lustre

Vast

Job description

Together AI is looking for a Staff Engineer to design, operate, and scale multi-petabyte storage systems and Kubernetes-native storage operators for AI workloads. You will work on high-throughput, low-latency data paths, multi-terabyte caching architectures, and integration of cutting-edge storage technologies for AI training and inference.

You will collaborate on securing production-grade storage with strict multi-tenancy, QoS, and automated provisioning while optimizing costs across thousands