Use Prevalidated Agents

​

Introduction

ZEDEDA Edge Inference Services provides a library of prevalidated agents. Each agent is a ready-to-deploy AI application package that has been tested and validated on ZEDEDA edge infrastructure, bundling an inference runtime and configuration optimized for a specific edge use case. The agent downloads its model from Hugging Face when it first starts, so the target cluster needs internet access for the first start. Prevalidated agents are designed to shorten the path from model selection to production deployment by providing a tested starting point that you can deploy, customize, and monitor through the standard ZEDEDA toolchain.

​

Prerequisites

• You have a role that can submit deployment requests: Organization Admin, MLOps Admin, or MLOps Engineer in your organization, or the Super Admin system role. See Organizations, Users, and Roles.

​

Available Prevalidated Agents

​

Autonomous Retail Billing Agent

The Autonomous Retail Billing Agent uses computer vision to identify items at a point-of-sale location and calculate the total bill without manual scanning. The agent runs inference at the edge to keep transaction data local and minimize network dependency. Key capabilities:

• Real-time product recognition from camera input.
• Billing calculation without network round-trips.
• Configurable product catalog and price mapping.

​

Faulty PCB Detection Agent

The Faulty PCB Detection Agent uses visual inspection to identify defects on printed circuit boards (PCBs) during manufacturing or quality control processes. The agent runs on nodes located near the production line to provide low-latency inspection results. Key capabilities:

• Defect classification across common PCB fault categories.
• High-throughput inspection suitable for production-line speeds.
• On-device inference with no cloud dependency during operation.

​

Deploy a Prevalidated Agent

Prevalidated agents are deployed through the standard application deployment workflow in ZEDEDA Edge Inference Services.

1. In the bottom navigation bar, select Inference.
2. Browse the available agents in your organization’s catalog.
3. Select the agent you want to deploy.
4. Follow the deployment wizard to specify the target ZEDEDA Edge Kubernetes Service cluster or cluster group, configure agent parameters, and submit the deployment request. You cannot target an individual edge node.

Deploying creates a deployment request with status Pending. Organization Admin or MLOps Admin in that organization must approve it; users cannot approve their own requests. For details on the deployment approval process, see Approve and Validate Models.

​

Monitor a Prevalidated Agent

After deployment, monitor the agent using the standard ZEDEDA monitoring tools:

• Infrastructure > Operations > Jobs: view deployment jobs and import tasks associated with the agent.
• Evaluation > Benchmarks: run benchmarks on the agent’s underlying model to measure performance on target hardware.

​

In This Section

• Edge Inference Services Overview
• Inference Dashboard
• Inference Quick Start
• Access and Navigate Edge Inference Services
• Organizations, Users, and Roles
• Manage Providers
  • Manage Providers Using Jupyter Notebook Integration
• Manage Models
  • Manage Models Using Jupyter Notebook Integration
  • Benchmark Models and Manage Device Pools
  • Monitor Benchmark Runs and Interpret Results
• Submit Model Deployment Requests
• Approve and Validate Models
• Monitor Inference Jobs
• Use ZEIA to Query and Operate Your Fleet
• Use Prevalidated Agents
  • Use the Faulty Printed Circuit Board Detection Agent
  • Use the Autonomous Retail Billing Agent
• Inference Troubleshooting