AI × AEC · CURATED INSIGHT
Bayer Crop Science Deploys Multi-Agent Decision Engine on Amazon Bedrock AgentCore
Bayer Crop Science built the Agentic Decision Engine (ADE) using Amazon Bedrock AgentCore to automate ML workflow operations, resolving approximately 80 percent of user inquiries without engineer intervention.
Bayer Crop Science introduced the Agentic Decision Engine (ADE) to help researchers and data scientists query machine learning workflows. Built on Amazon Bedrock AgentCore and orchestrated with LangGraph, ADE coordinates 22 specialist agents under a supervisor agent to extract context across logs, APIs, and metadata stores.
Key takeaways
- ADE runs 22 specialist agents to retrieve workflow status and telemetry, reducing diagnostic time from over an hour to under 10 minutes.
- The system resolves about 80 percent of queries without engineer intervention, while caching stable prompt prefixes reduced input token costs by 20 to 30 percent.
Source：AWS Machine Learning Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human.