About
KASI AI Hub is where AI and advanced-computing activity at KASI (Korea Astronomy and Space Science Institute) comes together — papers, computing resources, tools, and people, all in one place. It’s built and operated by the Center for AI & Advanced Computing (CAAC), a research center within KASI’s Observation Infrastructure Division. CAAC builds and operates the AI and computing infrastructure that powers modern astronomical research — from petabyte-scale data archives to domain-specific AI models that work alongside scientists.
Vision
The core hub leading future astronomical research through AI-based advanced computing.
Our goal is to generate transformative research outcomes by building high-performance AI computing infrastructure and developing domain-specialized AI companion scientists.
Core Capabilities
| Capability | Description |
|---|---|
| Astronomical data management | Systematic storage, curation, and utilization of large-scale astronomical datasets |
| Advanced AI & context engineering | Deployment and orchestration of large-scale AI models tailored to astronomical problems |
| AI/ML infrastructure | Building scalable, heterogeneous systems capable of handling diverse AI/ML workloads |
| Domain-specific AI agents | Developing AI astronomer systems that can propose hypotheses, plan observations, and verify results |
Key Activities
AI-ready Data Storage & Access
- Storage and management of unstructured (image) and structured (tabular) astronomical data
- External data sharing and open-access support
- Expanded VO (Virtual Observatory) and API interfaces for efficient programmatic access
Research Intelligence
- Large Language Model (LLM) infrastructure for astronomical AI agent environments
- Frontend and workflow development support for deploying in-house AI models as services
Hardware & Software Infrastructure
- Hierarchical, scalable on-premises high-performance computing (HPC/GPU accelerator) infrastructure
- Flexible integration with external large-scale computing facilities
- Organic, adaptive infrastructure for diverse AI/ML workloads
AI Astronomer Ecosystem
- AI-based system covering the full cycle: hypothesis generation → observation planning → result verification
- Open testbed for community access to AI astronomer tools
Recent Achievements
- Open-source cloud computing infrastructure — deployed and operating an open-source based cloud HPC environment for astronomical research
- KASI Science Data Portal — development of a unified platform for sharing and utilizing astronomical research data
- Integrated analysis environment — large-scale research data analysis and application development platform
- Modern astronomical archive system — archive infrastructure for contemporary astronomy records and data
Staff
The center currently comprises 4 full-time researchers embedded within KASI’s Observation Infrastructure Division.
Contact
For collaboration inquiries or access to computing resources, please reach out through the KASI research portal or contact the center leadership directly.