Save millions on inference by building a model router

Our speaker this month is Aria. He's going to teach you how to send every request to your AI agent exactly where it belongs instead of sending them all to the biggest, slowest, and most expensive models.

What we'll cover:
- Why "just use the best model" quietly costs you ~7x more than it should
- Why you can't just ask every model and keep the best answer
- The four ways to build a router, and which one you can finish today
- How this is the same idea as a load balancer
- How to catch a bad answer before a user sees it

You'll actually build something:
- We start with the code already running on your screen
- You'll write your own routing rule and watch the decisions change as you type
What you'll walk away with:
- A working router you can read end to end in one sitting
- The bigger pattern. This same shape routes concept supports many applications: tickets, alerts, tool calls, search, etc... Not just routing AI models

Who it's for:
- Anyone who has used AI before. No engineering, or machine learning background needed.
Bring:
A laptop or just a phone. That's it. No account, no API key, no sign up needed.

Run of show:

• 6:00 to 6:45, show up, grab a beer or some food (not sponsored), say hi
• 6:45 to 7:15, Aria's presentation
• 7:15 to 8:00, Hang out, ask more questions

Free to attend, and a bar tab will be sponsored by Obvious.

Matthew & Neel