Building Serverless AI powered Applications on AWS
summary
Two working demos show concrete results: a podcast RSS→MP3 per paragraph pipeline that parses an RSS feed, calls Amazon Polly to synthesize each paragraph, stores MP3s in S3 and assembles an iTunes-ready XML (deployed with Serverless framework); and a mobile web app hosted on S3 using Cognito tokens that uploads photos to S3, calls Amazon Rekognition for labels/faces and Amazon Polly to read results aloud (demo runs live, occasional browser audio hiccups). Deployment examples use AWS CLI scripting, Lambda, Step Functions, Route53 and mention model sizes (ResNet ~270MB, Lambda load limit ~500MB) and a common image labels with ~70% confidence threshold.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 1:35EventDriven
- 4:20Service
- 5:50AI
- 9:42Polly
- 13:33Demo
- 14:13Code
- 19:19Podcast
- 22:28Architecture
- 22:45Image Analysis
- 23:34Recognition API
- 24:55Confidence Score
- 25:48How to use it
- 26:48Demonstration
- 28:07Step Workflow
- 32:42Live Demo
- 36:37Deep Learning MI
- 42:23Diabetes Detection
- 43:44Autonomous Driving Systems
- 45:05Resonant
- 47:33Questions
related talks
01h 18mBuilding applications with Serverless Framework ...
25mBuilding serverless applications with Amazon S3
49mServerless Architectural Patterns and Best Practices
41mServerless LLM with AWS Lambda and LangChain: A Revolution in Application Development
39mGenerative AI Solution with AWS Serverless and Bedrock
20mThe Serverless First Mindset
42mNo Server Is Better Than No Server: Building Blocks for Serverless Architectures
38mBuilding Practical, Cost-Efficient GenAI Solutions Using Serverless