Skip to main content
Loading navigation
Back to challenges

Autonomous Process Operations Team with CrewAI

Process automation facilities suffer from critical skill shortages and complex shift transitions across operations, maintenance, and safety departments. You will create a multi-agent process operations squad using CrewAI where specialized agents (OperatorAgent, MaintenanceAgent, SafetyOfficer) collaborate to resolve chemical process drift. The squad must reach consensus on corrective actions within 3 interaction turns.

EngineeringHosted by Vera
Sign in free to start

What you are building

Orchestrate a role-playing team of collaborative AI agents using CrewAI to automate chemical refinery shift handovers and process drift management.

How work is evaluated

Evaluation

Evaluates multi-agent consensus output quality and task turn count.

Datasets

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

How submissions are scored

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

unanimous_approval_check

Verifies that Operator, Maintenance, and Safety agents all explicitly approved the final plan.

Binary check

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

Dimension 2

agent_interaction_turns

Number of round-trip agent messages to reach consensus • target: 2 • range: 1-6

Binary check

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

Learning goals

  • Design multi-agent role configurations in CrewAI with specific goals, backstories, and tools

  • Implement sequential and hierarchical task delegation among process operations agents

  • Integrate simulated chemical reactor telemetry analysis tools into CrewAI task executions

  • Evaluate multi-agent consensus reliability and resolution efficiency

Resources

Dataset notes

Continuous batch chemical reactor logs with simulated exothermic drift incidents.

Challenge setup

Evaluates multi-agent consensus output quality and task turn count.

Challenge input

JSON containing reactor_id, pressure_psi, temp_c, and catalyst_flow_rate

CrewAI

Required framework assignment.

Evaluated output

JSON with consensus_action, operator_approval, maintenance_approval, safety_approval, and total_turns

Evaluation checks
  • Verifies that Operator, Maintenance, and Safety agents all explicitly approved the final plan.
  • Number of round-trip agent messages to reach consensus • target: 2 • range: 1-6
Configured evaluation evidence
  • Benchmark: Multi-Agent Industrial Consensus Benchmark
  • Agent Interaction Turns target: 2
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • CrewAI · Required
  • crewAI · Optional

Evaluation contract

  • unanimous_approval_check · Weight 1
  • agent_interaction_turns · Weight 1

Metrics

  • agent_interaction_turns · Target 2

Benchmarks

  • Multi-Agent Industrial Consensus Benchmark
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing reactor_id, pressure_psi, temp_c, and catalyst_flow_rate

    Output: JSON with consensus_action, operator_approval, maintenance_approval, safety_approval, and total_turns

    Execution harness: Python

Evaluation tools

  • CrewAI · Multi-Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Autonomous Process Operations Team with CrewAI