Deskripsi Pekerjaan
HyperScale Systems is searching for a Senior AI Platform Engineer to architect resilient, high-performance distributed infrastructure for frontier-scale deep learning models.
You will own the lifecycle of our high-density GPU computing clusters, build high-throughput data ingestion pipelines, and deliver developer tooling for core research teams. This role requires deep systems-level knowledge, expertise in modern orchestration frameworks, and a continuous delivery mindset.
Tanggung Jawab
- Architect and maintain multi-cloud Kubernetes clusters running distributed training jobs across thousands of GPUs.
- Profile, diagnose, and resolve network, I/O, and compute bottlenecks across model training workflows.
- Automate infrastructure provisioning using Terraform and continuous delivery pipelines via ArgoCD.
- Build resilient observability, metrics aggregation, and auto-remediation tools using Prometheus, Grafana, and OpenTelemetry.
- Enforce least-privilege access, audit controls, and platform network isolation policies.
- Partner with ML researchers to package, deploy, and scale proprietary deep learning frameworks efficiently.
Kualifikasi
- 5+ years of production experience managing cloud-native infrastructure on AWS, GCP, or bare-metal setups.
- 3+ years operating large-scale Kubernetes workloads, including custom controllers and operators.
- Demonstrated experience provisioning and tuning multi-node GPU clusters using NVIDIA container runtimes and CUDA.
- Strong programming proficiency in Go, Python, or Rust for systems automation and internal tooling.
- Practical expertise with distributed storage solutions such as Lustre, Ceph, or high-throughput S3 integrations.
- Deep understanding of Linux kernel internals, network fabrics (InfiniBand/RoCE), and storage subsystem tuning.