Services
MUSCULAR Cluster provides three integrated service models to support research workloads with strong governance and scalable collaboration.
1) Private Compute Space (Container-Based)
Each user/project receives a private container environment with separation from others. Researchers can specify requested GPU, RAM, storage, and expected duration. This enables predictable allocation, isolation, and accountability.
2) Federated Learning Platform as a Service
Collaborative AI training across organizations without moving data from its original source. Suitable for sensitive domains such as medical data and regulated datasets, enabling multi-institution model development with privacy-aware practice.
3) International Federated Research Network
A cross-border collaboration capability to connect with international partners (e.g., Taiwan and UK) and to scale usage across Mahidol’s 11 network hospitals for joint medical AI and translational research.
How allocation works
- Submit request (resource + timeframe + project details)
- Review for feasibility and governance constraints
- Provision isolated workspace
- Onboarding guidance and operational support
Typical workload examples
- Medical imaging analytics and model training
- Large-scale data preprocessing and feature engineering
- Federated learning across hospitals or institutions
- Simulation / computational research requiring accelerators