HPC Expert ClearML & Confidential Computing

Key Responsibilities
Administer ClearML Server, ClearML Agents, execution queues, projects, users, roles, experiment tracking, pipelines, datasets, artifacts, and model registry.
Configure ClearML Agents on CPU and GPU worker nodes and integrate them with Slurm, PBS Professional, Kubernetes, or equivalent HPC execution platforms.
Support GPU-based AI/ML workloads using NVIDIA drivers, CUDA, NCCL, UCX, and containerized environments.
Maintain secure container execution using Apptainer/Singularity, Docker, Enroot, Pyxis, or Kubernetes.
Implement confidential-computing controls using technologies such as AMD SEV-SNP, Intel TDX, NVIDIA Confidential Computing, Secure Boot, TPM, and remote attestation where supported.
Integrate authentication, RBAC, TLS certificates, secrets management, and audit controls for ClearML and confidential workloads.
Monitor ClearML services, agents, queues, GPU utilization, task failures, scheduler integration, and platform health.
Automate deployment, configuration, monitoring, and troubleshooting using Python, Bash, Ansible, and Git.

Essential Skills
Solid Linux administration: RHEL, SLES, Rocky Linux, or Ubuntu.




Hands-on ClearML administration or robust experience with a comparable MLOps platform.
Experience with HPC schedulers: Slurm, PBS Skilled, or LSF.
GPU platform knowledge: NVIDIA drivers, CUDA, GPU monitoring, distributed training basics.
Containers: Apptainer/Singularity, Docker, Enroot, Pyxis, or Kubernetes.
Scripting and automation: Python, Bash, Ansible, Git.
Security fundamentals: RBAC, IAM, TLS, certificates, secrets management, secure boot, and audit logging.
Understanding of confidential-computing concepts: TEE, encrypted memory, TPM, remote attestation, AMD SEV-SNP, Intel TDX, or equivalent technologies.

MANDATE SKILLS:

HPC infrastructure experience,

(RHEL/SLES/Ubuntu)

HPC cluster management

Slurm or PBS Pro

InfiniBand networking,

HPE Performance Cluster Manager (HPCM) or Bright Cluster Manager,

Automation using Bash, Python, or Ansible

HPE ProLiant Gen10/Gen11 servers

Grafana/Prometheus monitoring,

Kubernetes, Docker/Singularity,

Parallel file systems such as Lustre or GPFS

PowerShell scripting and ITIL processes is a strong advantage

📌 Hpc Administrator Bengaluru
🏢 BaseBiz
📍 Bengaluru