Skip to main content
Back to challenges
Challenge

Predictive Downtime Notification Function with OpenAI Agents SDK

Automating predictive maintenance requires lightweight function calling agents that evaluate incoming machine alerts and trigger immediate technician notifications. In this challenge, developers will utilize the OpenAI Agents SDK to build a notification dispatch tool that filters baseline operational telemetry from true severe breakdown signals.

EngineeringHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Develop an event notification agent using OpenAI Agents SDK to evaluate machine fault alerts and trigger alert webhooks.

Delivery guide

How work is evaluated

Evaluation

Evaluates whether the agent correctly invokes the send_notification tool only for high severity alerts.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

Alert Trigger Verification

Critical alerts must invoke the notification tool

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

Precision Score

Correct tool invocation without false positive triggers on normal telemetry • target: 1 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Define agent tools using OpenAI Agents SDK `@function_tool` decorators

  • Filter non-critical telemetry warnings to avoid technician alert fatigue

  • Format high-severity notification payloads with machine location and estimated TTD

  • Execute simulated Slack/SMS notification function calls

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic breakdown alerts labeled with required dispatch urgency.

Challenge setup

Evaluates whether the agent correctly invokes the send_notification tool only for high severity alerts.

Challenge input

JSON machine_status object

OpenAI Agents SDK

Assigned framework

Evaluated output

JSON object with tool_invoked (boolean), recipient_channel (string)

Evaluation checks
  • Critical alerts must invoke the notification tool
  • Correct tool invocation without false positive triggers on normal telemetry • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: Function Calling Verification Benchmark
  • Precision Score target: 1
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • OpenAI · Required
  • Azure OpenAI Service · Optional
  • OpenAI o3 · Optional

Evaluation contract

  • Alert Trigger Verification · Weight 1
  • Precision Score · Weight 1

Metrics

  • Precision Score · Target 1

Benchmarks

  • Function Calling Verification Benchmark
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON machine_status object

    Output: JSON object with tool_invoked (boolean), recipient_channel (string)

    Execution harness: Python

Evaluation tools

  • OpenAI Agents SDK · Agent framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Predictive Downtime Notification Function with OpenAI Agents SDK