Back to evaluations
Draft evaluation

Build Wastewater Aeration Anomaly Agent with Vercel AI SDK and Gemini 2.5 Flash — evaluation

Evaluates the AI SDK agent's ability to classify telemetry anomalies and generate valid setpoint adjustments.

Evaluation type
task based
Challenge
Build Wastewater Aeration Anomaly Agent with Vercel AI SDK and Gemini 2.5 Flash
Difficulty
Intermediate
Rigor
Not declared

The author has not specified a rigor level.

Open linked challenge

Evaluation overview

How the linked challenge is judged: tasks, benchmarks, and criteria count.

Tasks
1
Benchmarks
0
Criteria
0

Task templates

Inputs and expected outputs.

Task 1

aeration_telemetry_eval

Tests anomaly detection accuracy on simulated wastewater aeration basin readings.

Input format

JSON containing telemetry arrays (DO, air flow, blower current)

Output format

JSON with anomaly_detected boolean, severity, and suggested_setpoint