Skip to main content
Back to challenges
Challenge

Google ADK Multifamily Property IoT Resilience Scoring

Commercial property underwriters struggle to correlate IoT water leak detectors and fire suppression investments with actual loss reduction. Implement a Google ADK agent that integrates building sensor telematics with property loss records to produce calibrated risk discount factors.

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.

Process IoT time-series data and building inspection imagery using Google ADK to re-rate property premiums.

Delivery guide

How work is evaluated

Evaluation

Evaluates accuracy of premium discount factors generated from property IoT telematics streams.

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

uptime_threshold_test

Ensures full discount is awarded only when uptime exceeds 98%

Binary check

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

Dimension 2

discount_accuracy_mae

Mean absolute error on computed premium discount percentage • target: 0.2 • range: 0-2

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

  • Build a Google ADK multimodal agent for parsing IoT sensor streams and building photos

  • Quantify expected loss mitigation percentages for automatic shutoff valves and smart fire alarms

  • Generate calibrated policy rate discount schedules based on real-time sensor uptime

Resources and assets

Reference links and supporting material

Dataset notes

Time-series water meter telemetry and boiler temperature logs from 100 multifamily residential complexes.

Challenge setup

Evaluates accuracy of premium discount factors generated from property IoT telematics streams.

Challenge input

JSON containing property_id, iot_devices_active, and 30_day_uptime_pct

Google ADK

Google's agent framework for multimodal and sensor data processing

Evaluated output

JSON with discount_credit_pct, expected_water_loss_reduction, premium_savings_usd

Evaluation checks
  • Ensures full discount is awarded only when uptime exceeds 98%
  • Mean absolute error on computed premium discount percentage • target: 0.2 • range: 0-2
Configured evaluation evidence
  • Benchmark: IoT-Underwriting-Bench
  • Discount Accuracy Mae target: 0.2
  • 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
  • Google · Required
  • Google ADK · Optional
  • fal.ai · Optional

Evaluation contract

  • uptime_threshold_test · Weight 1
  • discount_accuracy_mae · Weight 1

Metrics

  • discount_accuracy_mae · Target 0.2

Benchmarks

  • IoT-Underwriting-Bench
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing property_id, iot_devices_active, and 30_day_uptime_pct

    Output: JSON with discount_credit_pct, expected_water_loss_reduction, premium_savings_usd

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent SDK Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Google ADK Multifamily Property IoT Resilience Scoring