Government

[Seoul Bio Hub] Building an Integrated Platform to Support AI Development by Bio Startups

Seoul Bio Hub built an integrated AI development platform so that medical and bio startups can easily develop and serve AI models using GPU resources, data, and development environments. This created a bio AI support foundation that covers everything from computing resources for resident companies to AI development and deployment.

CheetahMLOps

An AI Development and Service Environment for Bio Startups

To support the growth of medical and bio startups, Seoul Bio Hub runs computing resource and data sharing programs and aimed to expand its foundation for bio AI development. This required an environment where resident companies could efficiently use infrastructure such as GPUs and easily develop and deploy AI models through to validation services.

Securing HPC Infrastructure Operations and a Framework for Bio AI Services

The key challenge was to efficiently operate the shared infrastructure provided to resident startups, while making it easy for anyone to start AI development and turn their models into services.

  • Infrastructure resource management: A framework to efficiently operate HPC resources such as GPUs provided to resident bio startups

  • Easy AI development environment: An environment where any startup can easily develop AI models with the libraries and frameworks they need

  • Model serving: A framework to quickly deploy developed AI models and connect them to real services

Building an Integrated Platform for Bio AI Development and Deployment

With Cheetah, we provided an environment for efficiently managing HPC infrastructure, together with a medical and bio AI development environment and a service framework based on one-click deployment.

  • HPC infrastructure operations: An environment to efficiently manage and distribute shared GPU resources on Kubernetes

  • Fast AI environment setup: Built-in AI libraries and frameworks allow medical and bio AI environments to be set up within one minute

  • One-click deployment: Developed models deploy quickly with automatic API generation, without building separate API servers

Faster AI Development and Service Rollout for Bio Startups

Seoul Bio Hub can now operate the HPC infrastructure it provides to resident companies more systematically, and startups can easily set up the environments they need to develop medical and bio AI models quickly. Developed AI models can also be managed by version and deployed easily, speeding up the transition to real validation services.

  • More efficient use of infrastructure: A framework to efficiently distribute and operate GPU resources for resident companies

  • Easier access to development environments: Easy setup of medical and bio AI environments reduces startups’ development burden

  • Faster service rollout: Model version control and automatic API-based deployment connect models to validation services

  • Are you curious about an enterprise AI integration platform? → Cheetah

Have more questions?

Contact us anytime to learn more about AIFRICA solutions.

Contact Us