What is Shoc Platform?
Shoc Platform is a serverless platform designed to simplify the execution of machine learning and high-performance computing workloads. It eliminates the need for infrastructure management by automatically packaging code into containers, scheduling jobs, and optimizing resource allocation across connected clusters.
The platform supports diverse workloads including HPC with OpenMPI and ML training with frameworks like TensorFlow and PyTorch. It provides integrated secrets management, a unified container registry, and enterprise-grade security with role-based access control, allowing users to focus on their code while the platform handles the underlying complexity.
Features
- Seamless Containerization: Automatically packages code into containers without manual Dockerfiles
- Advanced Job Scheduling: Queues, schedules, and optimizes resource allocation for HPC and ML workloads
- Cluster-Aware Execution: Intelligently distributes jobs across multiple clusters based on resource availability
- Integrated Secrets Management: Securely stores and injects credentials at workspace and user levels
- Unified Container Registry: Built-in registry or connect your own with access control enforcement
- Enterprise-Grade Security: Fine-grained role-based access control for resource isolation
Use Cases
- Running machine learning training jobs with TensorFlow or PyTorch
- Executing high-performance computing workloads with OpenMPI
- Managing distributed computing across multiple clusters
- Securely handling sensitive credentials in computational jobs
- Simplifying container management for computational workloads
