Engrain by ByteRush

AIF.01 3 concepts

Streaming UIs

Stream model output into the UI smoothly, accessibly and cancellably.

  1. AIF.01.01How LLM streaming works on the wireCoreMid-levelBeing written
  2. AIF.01.02Parsing and rendering token streamsCoreMid-levelBeing written
  3. AIF.01.03Latency, errors and accessibility of streamingMid-levelBeing written

Interview questions in this module

  1. How would you stream an LLM response into a React UI?
  2. Why should the browser not call the LLM provider directly?
  3. How do you handle a stream that fails halfway?
  4. How do you keep auto-scroll from fighting a user who scrolls up?
  5. How do you make streaming text accessible to screen-reader users?
  6. How do you render markdown that arrives token by token without re-parsing everything?