Simple AI Inference

Get Started

AI Inference Service that Provide AI Foundation Models

Simple AI Inference is a fully managed service delivering diverse, reliable AI foundation models. It provides both private and public endpoints with seamless access to AI models from inside and outside Samsung Cloud Plarform. It is compatible with OpenAI API and can be easily linked to existing development environments or frameworks, improving AI application development productivity.

Overview

  • Easy to Use AI Models

    Simple AI Inference is a serverless service that does not require infrastructure construction, allowing multiple types of AI models to be called through a single API. In addition, various latest open-source models can be used immediately after application, greatly improving the convenience of development.

  • Reliable Service Environment

    Limit Tokens Per Minute(TPM) and Requests Per Minute(RPM) to maintain consistent service availability in the event of a sudden traffic surge. It keeps services stable by precisely controlling data traffic even when processing large amounts of data.

  • Robust Security

    Customer data is never used for model training, ensuring it remains safe from leaks and strictly protected within Samsung Cloud Platform’s secure environment.

  • Efficient Cost Management

    It is charged based on the actual number of Input/Cached/Output tokens used. This allows flexible and highly efficient cost management tailored to your service scale and traffic.

01

04

Service Architecture

User → SCP Region's Virtual Server, Console → SCP Region's SCP Console SCP Region Virtual Server → Simple AI Inference: Provide Endpoint(Private Endpoint, Public Endpoint) → Inference(LLM Model, Safety Guard, Manager)
SCP Console → Create/Delete/Modify → Inference

Key Features

  • Easily view LLM models
    1. LLM model catalog : Easily view model-specific features and key use cases
    2. Open source models : Supports multiple open-source models that can be easily viewed through the console
  • Flexible model usage
    1. Model request : Requested models are available to all users within the same account
    2. Provide serverless service : Requested models are available immediately via API
    3. Provide public/private endpoints : Provides public endpoints for external access and private endpoints for Samsung Cloud Platform resources
  • Reliable service operations
    1. Traffic control : Enhances operational stability through TPM and RPM rate limiting
  • Robust security
    1. Safety guard : Features designed to ensure the safety, security, and reliability of AI services
    2. Enterprise security : Guarantees absolute data protection by ensuring customer data is never used for external model training

Let’s talk

Whether you’re looking for a specific business solution or just need some questions answered, we’re here to help

Get in Touch