Orchestrating Critical Mineral Supply Chain Resilience
As global trade accounts for over 60% of critical mineral demand, supply chain managers face unprecedented volatility from regulatory shifts like Chile's Codelco-SQM deal and the US NEPA overhaul. This challenge requires building an autonomous multi-agent system designed to monitor, analyze, and mitigate supply chain risks. You will leverage the Grok-2 model for reasoning and the CrewAI framework to orchestrate specialized agents focusing on geopolitical risk, regulatory compliance, and market logistics. Participants will implement a Retrieval-Augmented Generation (RAG) architecture using Pinecone to store and query real-time commodity data and policy updates. The system must process disparate data sources—ranging from mining project capital intensity reports to environmental impact legislative changes—to produce actionable risk mitigation strategies. The final output will be a resilient supply chain dashboard that predicts potential bottlenecks in battery metal procurement.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop a multi-agent AI system using CrewAI and Grok-2 to synthesize critical mineral market data and regulatory changes into dynamic supply chain risk assessments.
How work is evaluated
Evaluation focuses on agent autonomy, retrieval accuracy from Pinecone, and the logical consistency of the Grok-2 generated risk reports.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Master agentic workflows by defining specific roles, goals, and backstories for CrewAI agents tailored to mining logistics.
Implement a Pinecone vector database index to manage high-dimensional embeddings of mineral trade reports and NEPA legal documents.
Design a recursive task hierarchy where a 'Geopolitical Analyst' agent feeds findings into a 'Logistics Optimizer' agent.
Build a custom toolset in Python that allows agents to query external commodity APIs (e.g., LME, Quandl) for real-time battery metal pricing.
Orchestrate a Grok-2-powered reasoning loop that evaluates the impact of 'Progressive Design-Build' legislation on current infrastructure timelines.
Deploy a Streamlit-based monitoring interface to visualize agent communication logs and final risk scores.
Optimize agent performance by tuning temperature settings and top-p sampling for Grok-2 to ensure analytical precision in risk forecasting.
Integrate structured output parsing to convert agent findings into standardized JSON risk reports for downstream ERP systems.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
Evaluation focuses on agent autonomy, retrieval accuracy from Pinecone, and the logical consistency of the Grok-2 generated risk reports.
Challenge input
JSON containing a specific news headline and current inventory data
Agent execution
No agent tool is listed for this challenge.
Evaluated output
JSON risk assessment with probability, impact score, and mitigation steps
- Verify that the Regulatory Agent successfully passes data to the Risk Analyst agent.
- Ensure the agent's output contains specific facts retrieved from the vector store.
- Reasoning Accuracy target: 8
- Python execution harness
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
No tool records are attached.
Evaluation contract
- Verify that the Regulatory Agent successfully passes data to the Risk Analyst agent.
- Ensure the agent's output contains specific facts retrieved from the vector store.
Metrics
- Reasoning Accuracy · Target 8
Workflow tasks
- Task 1
Input: JSON containing a specific news headline and current inventory data
Output: JSON risk assessment with probability, impact score, and mitigation steps
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.