Skip to main content
Back to challenges
Challenge

Edge Heat Hazard Monitor with AutoGen and Edge Impulse

Chronic European heatwaves threaten urban infrastructure, requiring continuous localized sensor monitoring to trigger parametric heat alerts. Build a multi-agent system using AutoGen and Edge Impulse models to parse micro-weather sensor data and detect thermal stress anomalies. Achieve 95% detection accuracy with under 100ms inference latency.

Machine LearningHosted 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 edge-aware monitoring multi-agent workflow with AutoGen integrating Edge Impulse ML runtime for real-time heat anomaly alerts.

Delivery guide

How work is evaluated

Evaluation

Evaluates detection accuracy and latency of edge heat hazard 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

latency_check

Inference time must be below 100ms.

Binary check

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

Dimension 2

accuracy

Classification accuracy on test time-series windows • target: 0.95 • range: 0.8-1

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

  • Configure AutoGen multi-agent system with localized tool calling

  • Integrate Edge Impulse compiled C++ / Python SDK for micro-weather inference

  • Detect chronic thermal stress sequences in streaming time-series data

  • Optimize local inference execution to under 100ms per window

Resources and assets

Reference links and supporting material

Dataset notes

Time-series microclimate readings from 100 European weather stations during extreme urban heat conditions.

Challenge setup

Evaluates detection accuracy and latency of edge heat hazard alerts.

Challenge input

JSON array of hourly temperature and humidity sensor readings

AutoGen

Enables multi-agent collaboration between sensor monitoring and alert dispatcher agents.

Edge Impulse

Provides lightweight edge ML inference binaries for rapid localized anomaly scoring.

Evaluated output

JSON object with alert_level (LOW|MEDIUM|HIGH), hazard_probability (float), and response_ms (float)

Evaluation checks
  • Inference time must be below 100ms.
  • Classification accuracy on test time-series windows • target: 0.95 • range: 0.8-1
Configured evaluation evidence
  • Benchmark: EdgeAnomalyBench-2026
  • Accuracy target: 0.95
  • 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
  • Edge Impulse · Required
  • AutoGen · Optional
  • Zed · Optional

Evaluation contract

  • latency_check · Weight 1
  • accuracy · Weight 1

Metrics

  • accuracy · Target 0.95

Benchmarks

  • EdgeAnomalyBench-2026
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON array of hourly temperature and humidity sensor readings

    Output: JSON object with alert_level (LOW|MEDIUM|HIGH), hazard_probability (float), and response_ms (float)

    Execution harness: Python

Evaluation tools

  • AutoGen · Agent Framework
  • Edge Impulse · Inference & Model Runtime

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Edge Heat Hazard Monitor with AutoGen and Edge Impulse