This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Staff Storage Platform Engineer (AI Storage) - Radian Arc based in France.
This is a Staff-level opportunity to shape the storage architecture powering large-scale GPU and AI infrastructure across edge and core environments.
You will design, build, and operate high-performance storage systems supporting training, fine-tuning, and distributed inference workloads.
The role covers hyperconverged, local NVMe, and disaggregated storage architectures, with a strong focus on throughput, latency, resilience, and cost efficiency.
You will work at the intersection of storage, GPUs, networking, Kubernetes, and AI platform engineering, ensuring storage never becomes a bottleneck for compute.
As the primary storage specialist, you will combine architectural ownership with hands-on engineering, troubleshooting, performance optimization, and deployment.
You will also define reusable standards, influence long-term platform direction, and mentor engineers across adjacent infrastructure domains.
This is an ideal environment for a senior storage expert who wants significant technical ownership and direct impact on next-generation AI infrastructure.
Accountabilities
- Storage architecture: Design scalable storage architectures for edge and core GPU deployments, covering hyperconverged platforms such as StorPool, local NVMe, and disaggregated systems such as VAST Data and Weka. Define reference architectures, reusable design patterns, fault domains, lifecycle strategies, and scaling approaches while balancing throughput, latency, resilience, data locality, operability, and cost.
- AI workload optimization: Optimize storage for distributed training, fine-tuning, and inference workloads, including large dataset ingestion, model artifact distribution, checkpointing, and high-concurrency access. Establish realistic performance baselines and ensure storage architecture aligns with actual GPU workload behavior.
- Distributed inference: Design storage architectures supporting inference platforms such as NVIDIA Dynamo, llm-d, or similar systems. Optimize model distribution, token-generation data paths, and KV-cache persistence and retrieval so infrastructure can scale efficiently across large GPU clusters without storage becoming a throughput or latency bottleneck.
- High-performance data paths: Engineer efficient storage-to-GPU data paths using technologies such as GPU Direct Storage, RDMA/RoCE, NVMe-oF, and SPDK. Investigate and tune performance across hardware, networking, operating systems, filesystems, storage layers, and distributed workloads.
- Platform integration: Integrate block, object, and shared file storage into Kubernetes and platform orchestration systems. Implement and maintain CSI integrations, support multi-tenant storage architectures, and define standards for storage integration across different deployment models.
- Distributed storage: Contribute to large-scale storage platforms, including S3-compatible object storage, distributed file systems, and block storage. Design systems with clear operational boundaries, resilience models, scaling paths, and reusable operating patterns across multi-cluster and multi-site environments.
- Performance and reliability: Lead storage benchmarking, capacity planning, performance investigations, incident response, and root-cause analysis. Establish measurable standards for throughput, latency consistency, recovery behavior, reliability, and operational maturity.
- Engineering delivery: Own storage initiatives end to end, from architecture and validation through production rollout. Validate BOMs, topology decisions, node profiles, and deployment assumptions while ensuring changes are introduced safely with minimal customer impact.
- Operational excellence: Improve storage observability, automation, runbooks, lifecycle management, and day-2 operations. Turn recurring incidents and operational pain points into durable engineering improvements and standardized practices.
- Technical leadership: Act as the primary storage design authority, influencing platform architecture and roadmap decisions across compute, networking, DevOps, infrastructure, and operations. Communicate technical trade-offs clearly, mentor adjacent engineers, and raise the organization’s expertise in AI storage.
Requirements
- Distributed storage expertise: Strong hands-on experience designing and operating distributed storage systems in high-performance computing, AI, GPU, or similarly demanding environments.
- AI infrastructure experience: Proven experience designing storage architectures for large-scale AI training, fine-tuning, or inference, including dataset distribution, model artifacts, checkpointing, and high-concurrency data access.
- AI storage knowledge: Deep understanding of how AI workload characteristics affect storage throughput, latency, concurrency, data locality, checkpoint recovery, and serving performance.
- Storage technologies: Hands-on experience with technologies such as Weka, VAST Data, StorPool, local NVMe, distributed filesystems, S3-compatible object storage, block storage, and/or comparable enterprise storage platforms.
- Linux and systems expertise: Strong knowledge of the Linux storage and I/O stack, storage hardware, NVMe devices, storage fabrics, and high-performance data paths.
- Kubernetes: Familiarity with Kubernetes storage integrations, particularly CSI, and experience integrating storage into containerized or orchestrated platforms.
- AI data paths: Practical knowledge of GPU Direct Storage, RDMA/RoCE, NVMe-oF, SPDK, and techniques for minimizing unnecessary data movement between storage and GPU compute.
- Distributed inference: Experience with storage requirements for inference orchestration and model-serving environments, including model distribution and KV-cache persistence or retrieval, is highly valuable.
- Troubleshooting: Ability to diagnose complex cross-layer issues involving storage hardware, networking, Linux kernels and I/O paths, filesystems, object/block storage, Kubernetes, and distributed workloads.
- Automation: Strong Python and/or Bash skills, with experience applying software engineering practices to infrastructure automation, validation, lifecycle management, and operational tooling.
- Observability: Experience designing or operating storage observability systems and using metrics and telemetry to identify performance, reliability, and capacity issues.
- Technical leadership: Demonstrated ability to lead complex infrastructure initiatives, establish architectural standards, and influence multiple teams without relying on formal management authority.
- Systems thinking: Ability to balance performance, scalability, reliability, operability, deployment complexity, and cost when making architecture decisions.
- Communication and collaboration: Comfortable working with compute, networking, platform, DevOps, operations, deployment teams, vendors, and other technical stakeholders.
- Ownership: Able to combine Staff-level strategic thinking with hands-on execution, particularly in a lean or fast-scaling environment where processes and standards are still being established.
- Mentoring: Strong ability to share knowledge, guide engineers in adjacent domains, and raise the technical bar across the broader infrastructure organization.
Benefits
- Attractive compensation package aligned with your expertise and experience.
- Opportunity to play a foundational role in shaping a next-generation AI storage platform.
- Significant architectural ownership and direct influence over long-term infrastructure strategy.
- Exposure to cutting-edge GPU, AI inference, distributed storage, and high-performance data technologies.
- International and diverse working environment with strong flexibility.
- Remote-friendly work model across Europe.
- Opportunity to join a fast-growing scale-up with an ambitious technology mission.
- Broad cross-functional exposure across infrastructure, compute, networking, platform engineering, and operations.
- Strong career growth potential as the infrastructure organization expands.
- Inclusive environment committed to equal opportunity and professional development.
How Jobgether works:
We use an AI-powered matching process to ensure your application is reviewed quickly, objectively, and fairly against the role's core requirements. Our system identifies the top-fitting candidates, and this shortlist is then shared directly with the hiring company. The final decision and next steps (interviews, assessments) are managed by their internal team.
We appreciate your interest and wish you the best!
Data Privacy Notice: By submitting your application, you acknowledge that Jobgether will process your personal data to evaluate your candidacy and share relevant information with the hiring employer. This processing is based on legitimate interest and pre-contractual measures under applicable data protection laws (including GDPR). You may exercise your rights (access, rectification, erasure, objection) at any time.
#LI-CL1
