On-device models: WebGPU, WASM and built-in AI
Comes up in: Frontend system design, AI system design
This concept is being written. It gets the full five steps — explanation, predict, interview angle, practice with tests, and review cards. Here is exactly what it will cover, so you can study it elsewhere in the meantime.
What it covers
- transformers.js / ONNX Runtime Web
- WebGPU vs WASM backends
- model size and download cost
- running in a worker
- privacy benefits
- browser built-in AI APIs (status per browser — facts)
- good use cases (embeddings, classification, background removal) vs bad ones
Words you will know
What you will practise
- in-browser sentiment classifier in a worker
Interview questions this answers
- When would you run an ML model in the browser instead of on a server?
- What are the costs of on-device models for users on low-end phones?