diff --git a/workshops/open-source-ai-models.md b/workshops/open-source-ai-models.md
new file mode 100644
index 0000000..729deef
--- /dev/null
+++ b/workshops/open-source-ai-models.md
@@ -0,0 +1,46 @@
+# Open Source AI Models
+
+A workshop for folks who are ready to dive into some more advanced Defang features and want to host their own LLM models.
+
+---SLIDE---
+
+## Why host your own model
+
+1. It can be less expensive
+2. You can experiment with a wide variety of models
+3. Can fulfil legal requirements
+
+---SLIDE---
+
+## What is Ollama?
+
+[Ollama](https://ollama.com/) is a tool that helps you easily manage and run open source (or open weight) llms.
+
+---SLIDE---
+
+## What's the easiest way?
+
+1. Head to the [Defang Samples](https://defang.io/#samples) and use the 1-click deploy workflow for Ollama
+2. Edit your code (can use [github.dev](https://github.dev)) as an easy way to get started.
+3. Can use [CodeSpaces](https://github.com/features/codespaces) as a way to work with the Defang CLI without having to install it.
+
+---SLIDE---
+
+## Diving deeper...
+
+1. Install the Defang CLI
+2. Clone the repo
+3. Run `docker compose -f compose.dev.yaml`
+4. Update, and deploy:
+    a. `git push`
+    b. `defang compose up`
+
+
+---SLIDE---
+
+## Can I use my own domain?
+
+1. Yes! Let's grab a domain.
+2. Great, now let's added it to our `compose.yaml` file and redeploy (`defang compose up`)
+3. Now run `defang cert generate`
+4. Success!