AI Inference Server¶

Reduce the cost, effort and time to bring AI models to the shop floor and help to focus on your domain expertise

Do you see AI as a potential solution for your current challenges? But are you concerned about the high costs and effort to bring AI to the shop floor and maintain it? Are you interested in standardizing the execution of AI Models on the shop floor and integrate it with monitoring, debugging, and remote deployment capabilities?

AI Inference Server is the edge application to standardize AI model execution on Siemens Industrial Edge. The application eases data ingestion, orchestrates data traffic, and is compatible all powerful AI frameworks thanks to the embedded Python interpreter. It enables the AI model deployment as content to a standard edge app instead of custom container deployment for single models. AI finally meets Siemens Industrial Edge and AI Inference Server leverages the benefits of both.

Key Features and Benefits¶

• Key Features

  • Supports most popular AI frameworks compatible with Python
  • Orchestrates and controls AI model execution
  • Ability to run AI pipelines both with an older and a newer version of Python
  • Allows horizontal scaling of AI pipelines for optimal performance
  • Simplifies tasks such as input mapping (thanks to the integration with Databus and other Siemens Industrial Edge connectors), data ingestion, pipeline visualization
  • Monitoring and debugging of AI models using inference statistics, logging and image visualization
  • Contains pipeline version handling
  • Import models via UI or receive them remotely
  • Supports persistent data storage on the local device for each pipeline

• Benefits

  • Standardize the AI model execution using an AI-ready inference with the Edge ecosystem
  • Standardize the logging, monitoring, and debugging of AI models
  • Designed for MLOps Integration with AI Model Monitor

Available APIs¶