Core Values
AI proves its value in operation, not adoption.
Cheetah connects GPU resources, MLOps and LLMOps workloads, models, and AI services in one operational flow.
Unified AI Operations
Connect GPU management, AI jobs, model
evaluation, deployment, and inference
in one consistently managed flow.
Automation · Human Judgment
Automate repeatable operations by policy,
while people control approvals, exceptions,
and key decisions for the right balance.
Data-Driven Improvement
Analyze accumulated operational and
quality data to continuously improve
and optimize your AI services.
Features
Cheetah: Unified GPUOps · MLOps · LLMOps
Efficient GPU Operations
Run limited GPU resources
by policy and priority
Maximize utilization of limited GPUs with flexible policies.
Auto-allocate and reclaim by user and group to cut idle capacity.
Provision the right resources for each workload via the catalog.
Reorder jobs by priority to share GPUs efficiently.
Build fast, operate reliably
Faster, more efficient MLOps
for the entire AI lifecycle
Manage the AI lifecycle end to end, from data to operations.
Spin up a validated standard dev environment in under a minute.
Automate repetitive tasks like data preprocessing and training.
Deploy trained models in one click and run them reliably.
Flexible, reliable model operations
Reliable LLMOps
built for on-premises
Optimize LLMs easily with a GUI, even in air-gapped networks.
Compare and validate models with Korean benchmarks and leaderboards.
Deploy validated models to diverse serving engines with zero downtime.
Monitor model health and adapt flexibly to changing loads.
Policy-based AI governance
Unify enterprise security and
AI policies in one Gateway
Manage every AI service request and response in a single Gateway.
Control access and call volume; block PII and harmful content.
Contain overloads and failures with Rate Limit and Circuit Breaker.
Track per-request results and audit trails to govern operations.
Unified, visible operations
Unified management of AI
resources, systems, and users
See GPU and CPU status and usage in real time on the dashboard.
Protect assets with group-based permissions and resource control.
Prevent misuse with request- and approval-based allocation.
Trace all operations transparently for safety and accountability.
Strengths
Key Strengths
Architecture
CHEETAH Workflow
Reference
End-to-end ML & LLMOps,
from development to operations
[HIRA] 2026 Cloud Transition and Business Service Improvement Project
The Health Insurance Review and Assessment Service (HIRA) adopted Cheetah, an AI platform, to streamline internal operations and expand AI services for the public, building an environment that supports a wide range of AI services from model development to agent creation and service operations.
Explore more customer stories.
With limited ability to use external AI models, Celltrion selected Cheetah as its AI platform to drive AX across internal research and business operations. Together, we built an environment for phased GPU expansion and company-wide AI development and operations.
Ulsan Technopark built a shared AI platform on Cheetah to support AI development and validation for local manufacturers, creating an environment that efficiently allocates GPU resources to resident companies and supports them from AI development through to validation.
DB Insurance built an AI-OCR platform to establish a company-wide AI OCR development and operations environment. This improved the accuracy and operational efficiency of insurance document processing and created a foundation for managing AI OCR models reliably.
