Create a First AI Program with TypeScript and Serverless WebAssembly

The workshop begins with a few minutes of introductory material on WebAssembly, serverless functions, and AI inferences. Then we will dive into the process of setting up an environment (CodeSpaces, Docker container, or local).

From there, we will build a first program in TypeScript (for most attendees) or Rust or Python (for attendees already familiar with those languages). This program is a serverless function that we will compile to WebAssembly and test out locally. After a few iterations on this code, we'll move to the next big task.

Once the Hello World part is done, we'll use an LLM-backed inferencing service to create a prompt-driven application. We'll write a basic prompt in code, try it out, and then refine it.

For the more advanced students, we'll build a magic 8-ball style program backed by the LLM.

For new developers, the outcome will be writing a basic serverless function and then extending it on to use AI. Since we provide a pre-built Code Spaces environment, anyone with a GitHub account can be up and coding in no time without needing to do a lot of local configuration.