Skip to main content
Challenge

Build a Refinery Incident Repair Scheduler with Google ADK and Gemini

Unplanned refinery damage requires emergency mobilization and rapid repair sequencing to safely restore refining capacity. Build an agent system using Google Agent Development Kit (ADK) and Gemini 1.5 Pro to synthesize incident damage logs and coordinate replacement parts with less than 5% scheduling downtime conflict.

Business OperationsHosted by Vera
Challenge brief

What you are building

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

Develop an emergency refinery repair scheduler using Google ADK with Gemini multimodal capabilities.

Delivery guide

How work is evaluated

Evaluation

Evaluates emergency repair task plan generated by Google ADK agent against refinery safety rules.

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

purge_before_weld_check

Ensures hydrocarbon purging step is sequenced before mechanical valve replacement

Binary check

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

Dimension 2

safety_permit_completeness

Percentage of required safety permits identified • 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

  • Build incident dispatch workflows using Google Agent Development Kit (ADK)

  • Process damage photos and text reports using Gemini multimodal capabilities

  • Generate prioritized repair work packs for Kuwait refinery units

  • Integrate automated resource availability checks against contractor inventory

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic refinery maintenance damage reports and part replacement catalogs based on Mina Al-Ahmadi and Al-Zour repair operations.

Challenge setup

Evaluates emergency repair task plan generated by Google ADK agent against refinery safety rules.

Challenge input

JSON containing unit_id, damage_severity, replacement_parts_needed

Google ADK

Google Agent Development Kit integration

Gemini

Multimodal vision and text analysis

Evaluated output

JSON containing repair_sequence, estimated_hours, safety_permits_required

Evaluation checks
  • Ensures hydrocarbon purging step is sequenced before mechanical valve replacement
  • Percentage of required safety permits identified • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: Industrial Emergency Dispatch Bench
  • Safety Permit Completeness 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
  • Google Gemini · Required
  • Google · Optional
  • Google ADK · Optional

Evaluation contract

  • purge_before_weld_check · Weight 1
  • safety_permit_completeness · Weight 1

Metrics

  • safety_permit_completeness · Target 1

Benchmarks

  • Industrial Emergency Dispatch Bench
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing unit_id, damage_severity, replacement_parts_needed

    Output: JSON containing repair_sequence, estimated_hours, safety_permits_required

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent Framework
  • Gemini · Multimodal Model

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Refinery Incident Repair Scheduler with Google ADK and Gemini