Using AI to power a circular economy for aerospace
summary
Lifecycle Optimization for Aerospace is a serverless AWS platform that consolidates aircraft paperwork to improve reuse, repair, resale and recycling; it ingests ZIP uploads, runs Textract OCR, a SageMaker BERT classifier, Bedrock (Claude v2) prompts inside nested Step Functions, and stores metadata in OpenSearch, with per-customer accounts deployed by CDK/CICD. It processes ~300,000 documents per aircraft, shows a human sample (18 parts) taking 2.5 hours vs robot <30 minutes, added generative AI in five days, and reports ~$0.10 per Bedrock inference. Textract OCR Bedrock Claude V2 300,000 documents per aircraft
AI-written summary · about
chapters (7)
- 0:00Industry challenge overview
- 1:51The circular economy vision
- 5:26Platform architecture and AI
- 12:37Data security and compliance
- 16:01Implementation and real results
- 22:15Material tracking optimization
- 24:50Generative AI and future path
related talks
43mImplementing AI Document Processing at Scale: Syngenta
14mDriving Industrial Innovation with AI at Scale
38mBuilding Practical, Cost-Efficient GenAI Solutions Using Serverless
20mThe Transformative Role of AI in Aerospace
50mAzure AI, Power new possibilities for every organization
38mGenerative AI - Architectures and applications in depth
43mBuilding an AI-Ready Data Foundation
20mBuilding a secure end-to-end generative AI application in the cloud