GPU Cloud Storage Architect

Bitdeer Group; Singapore; On-site; SGD 90,000 - 130,000; Full time

Job summary

Bitdeer is seeking an experienced Storage Infrastructure Engineer to design, implement, and optimize distributed and parallel file systems for its GPU cloud. You will drive end-to-end storage integration, ensuring multi-tenant isolation, provisioning, and lifecycle management across a multi-region setup.

Responsibilities include tuning throughput/latency for large-scale AI workloads, managing golden images and CUDA drivers, and building monitoring and runbooks.

Qualifications

• 3+ years (Senior 6+) in storage engineering or platform infrastructure.
• Strong understanding of distributed file system internals — data / metadata separation, replication, consistency models.
• Proven experience integrating and operating distributed storage in production (Ceph, Lustre, GPFS / Spectrum Scale, BeeGFS, JuiceFS, MinIO).
• Performance tuning for high-throughput / parallel I/O; familiarity with NVMe, RDMA / RoCE storage networking, and caching is a strong plus.
• Strong Linux systems depth and automation skills (Python / Go, CI / CD).
• HPC / AI storage or multi-region storage experience a strong plus

Responsibilities

• Design and integrate distributed / parallel file systems into the GPU cloud, optimized for AI training / inference I/O patterns.
• Own end-to-end distributed-storage integration: provisioning, mounting, multi-tenant isolation, quota, and lifecycle within the platform / control plane.
• Tune storage throughput and latency for large-scale parallel access (dataset loading, checkpointing); benchmark across GPU SKUs and workloads.
• Architect multi-region storage: data locality, replication / consistency, durability (failure domains), and cross-region access.
• Own golden images, templates, GPU drivers / CUDA, and the container / image registry, including versioned release and multi-region distribution. (Secondary scope.)
• Build monitoring, capacity planning, and runbooks; eliminate single points of failure.
• Partner with Compute (delivery), Network (storage fabric / RDMA), and Control Plane (provisioning / quota) teams.

Skills

Distributed FS

Linux

Python

Go

CI/CD

NVMe

RDMA

RoCE

GPU drivers

CUDA

Container registry

Automation

Performance tuning

Multi-region storage

Tools

Ceph

Lustre

GPFS / Spectrum Scale

BeeGFS

JuiceFS

MinIO

Job description

Bitdeer is seeking an experienced Storage Infrastructure Engineer to design, implement, and optimize distributed and parallel file systems for its GPU cloud. You will drive end-to-end storage integration, ensuring multi-tenant isolation, provisioning, and lifecycle management across a multi-region setup.

Responsibilities include tuning throughput/latency for large-scale AI workloads, managing golden images and CUDA drivers, and building monitoring and runbooks.