HPC Engineer

About the role

AlphaGrep is seeking an experienced HPC / GPU Cluster Engineer to design, operate, and continuously optimize a large‑scale GPU compute environment for distributed model training and high‑performance workloads.

Key responsibilities

• Design, deploy, and maintain GPU hardware, interconnect fabric, storage layers, and scheduling systems.
• Develop and maintain tooling for monitoring, automation, diagnostics, and benchmarking of hundreds of accelerators.
• Troubleshoot performance bottlenecks across the full stack, from low‑level RDMA to application level.

Required profile

• Proven experience operating large GPU or HPC clusters in production.
• Hands‑on expertise with RDMA programming, InfiniBand/RoCEv2, GPUDirect technologies, and performance benchmarking.
• Strong Linux system administration skills, including networking, kernel/driver issues, and resource management.

Required skills

• RDMA verbs / libibverbs, memory registration, queue pairs, completion queues.
• InfiniBand, RoCEv2, GPUDirect RDMA, GPUDirect Storage, NCCL, CUDA drivers, OFED.
• Slurm scheduler installation, configuration, and job troubleshooting.
• High‑performance shared storage/file systems such as Lustre, BeeGFS, WEKA, VAST, DDN/ExaScaler.
• Python, Bash, and preferably C/C++ for automation and monitoring.

What we offer

• Trust‑based collaborative culture with open communication.
• Flat organization, regular team‑building events and off‑sites.
• Health and wellness benefits including gym subsidy, stocked kitchen, and generous vacation.