Building a Serverless Compute Platform with AWS and TypeScript
summary
A four‑person team builds a serverless compute platform in weeks not years, letting customers upload arbitrary TypeScript/JavaScript functions and invoke them with JSON events. Code and metadata live in S3 and DynamoDB, deployment uses AWS CDK, and runtime isolation uses vm2 + worker threads (via frets.js) with hard timeouts and killed workers to avoid event‑loop blocking. Execution runs in a separate AWS account with minimal permissions. Frontend uses Monaco editor plus ESLint and Prettier to prevent bad saves.
AI-written summary · about
chapters (8)
- 0:00Introduction and motivation
- 2:28Speaker introduction and EDI context
- 7:45Why TypeScript for full stack
- 11:17System architecture overview
- 15:27Security and isolation strategies
- 19:57Runtime sandboxing and worker threads
- 22:02Frontend and code editor setup
- 26:15Key takeaways and recommendations
related talks
27mShould I fire my DevOps already? Build your infrastructure using TypeScript
39mDevelopment productivity in a post serverless world
51mServerless - a backend thing that gives superpowers to frontend developers
44mServerless: Giving superpowers to frontend developers
20mThe Serverless First Mindset
41mYou Don’t Want to Build a Serverless Monolith
25mGetting into the Cloud with TypeScript
30mThinking Serverless: From User Request to Serverless Solution
