Engineering Excellence: Platforms and Architectures for Real-Time AI Analytics
summary
Claims that production real-time AI only works with platform engineering and demonstrates a concrete stack: event ingestion via Kafka, microservices scaled with Kubernetes and Docker, model hosting on AWS SageMaker/Bedrock, and storage in a vector database. Describes a fraud-detection pipeline that must handle millions of transactions per hour with autoscaling, anomaly detection (autoencoders or TensorFlow), and centralized observability and secrets/encryption; uses coffee-maker and plant-bot examples to show end-to-end triggers, orchestration, and monitoring.
AI-written summary · about
chapters (7)
- 0:00Introduction and speaker background
- 3:02AI and real-time analytics overview
- 6:34AI coffee maker example
- 14:38AI plant care system example
- 18:28Engineering excellence in AI
- 23:02Fraud detection system architecture
- 25:29Conclusion and future trends
related talks
16mOperationalizing AI in Finance: Real-Time Analytics
13mScaling AI Systems for High-Volume Transaction Processing
04mInterview
35mHigh-performance data foundation for AI success
13mBuild Real-Time Analytics Platforms for Financial Crime Detection
17mBuilding Resilient AI Pricing Platforms
50mThe tech behind modern AI/ML use cases in Gaming
30mEnterprise AI: How To Operationalize Data Analytics At Scale