Government

[Incheon Startup Park] Building the POOM Platform for Startup AI Development, Deployment, and Validation

Incheon Startup Park built the POOM platform so startups can go from using data to developing, deploying, converging, and validating AI models as services. To do this, it adopted Cheetah for an integrated AI development environment and Serengeti for a service convergence and validation environment.

CheetahLLMOpsMLOpsSerengeti CMPSerengeti PaaS

An Integrated Environment from AI Development to Validation

Incheon Startup Park needed an environment that connects data collection, AI development, deployment, and validation step by step, so startups could use data to discover new business opportunities and develop services. It also required a framework to reliably operate its HPC infrastructure while letting resident companies quickly set up and use the analytics and development environments they need.

Securing AI Infrastructure Operations and Model Development and Validation

The key challenge was to enable resident startups to manage data, develop and deploy AI models on shared infrastructure, and connect them to service validation.

  • HPC infrastructure operations: A foundation to reliably manage HPC resources and provide them systematically to resident companies

  • Data and development environments: A framework for registering and managing large-scale data and setting up tailored AI analytics and development environments for each company

  • Fast deployment and validation: An environment to quickly deploy developed AI models and connect them to service convergence and validation

Building the POOM Platform by Connecting the AI Portal and PaaS Portal

With Cheetah, we provided AI infrastructure management and a data and model development and deployment environment, and with Serengeti, we built an integrated environment for converging deployed models and APIs into services and validating them.

  • Cheetah-based AI development environment: Integrated support for HPC resource management, dataset registration and sharing, and tailored AI development environments

  • Fast model deployment: Shorter validation lead time with one-click deployment and automatic REST API endpoint creation

  • Serengeti-based service convergence: Supported service validation by connecting deployed model APIs and diverse features with drag and drop

A Foundation for AI Development and Validation Support for Startups

Incheon Startup Park now provides an integrated environment where resident companies can use its infrastructure to quickly develop and deploy AI models and carry them through to service validation. This effectively established an AI foundation that supports startups in developing new services and creating new business.

  • AI development and deployment environment: An environment where resident companies can quickly develop and deploy data-driven AI models

  • Converged validation environment: A foundation for validating services by connecting finished models with diverse APIs

  • Stronger support for startup growth: Support from AI model development to service launch expands opportunities for new business

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

Have more questions?

Contact us anytime to learn more about AIFRICA solutions.

Contact Us