The Eval Edition - Know before you ship

​How do you know your AI actually works?
We’re talking evals - the most underhyped part of building real-world AI.

​This edition of Two Tokens Friday is all about how to evaluate, test, and trust the AI you’re shipping.

​What’s on the Menu

​Tool Drops
Eval Playground – Build & run evals with ease. Plug in your prompts, datasets, and run side-by-side comparisons.
Structured Feedback Infra – What users feel vs. what models do. We’re bridging the gap.

​Headliner
A product leader from one of India’s top AI startups joins us to share how their team runs evals in production - from model changes to feature launches.

​Hands-On Fun
Red Team Roulette – Spot model failures and earn swag.
Build Your Own Eval Run – Try out our tools live and take your evals beyond spreadsheets.

​Snacks & Sharp Talk
No filler. Just honest conversations about what works, what fails, and how teams are keeping AI grounded in reality.

​Who It’s For

• ​Founders & PMs pushing AI to prod

• ​Engineers building evaluation infra

• ​Researchers tired of toy benchmarks

• ​Anyone who wants to move beyond "looks good"

​Bottom Line

​If you’ve ever asked “Is this model good enough?” this event is for you.

​Walk in curious. Walk out with a playbook.