We Taught AI to cook steak

“Make The Perfect Steak”

When we started Gambit, I made a pact: no more ordering in. Since steak is my favorite food, I've probably gone to nearly every butcher in NYC over the last few months.

A core part of our stack is AI, specifically computer vision. With RGB and thermal cameras, Gambit interprets what's happening in your pan and tracks how food is cooking in real time.

So, it should come as no surprise that I wanted to roll out a "Make the Perfect Steak" feature. Steak is expensive, tricky to nail, and the perfect hero use case for Gambit.

Building Steak AI

• Collected a large training set of steak images

• Labeled cut, marbling, surface color, and doneness stage

• Curated rules for size, thickness, doneness preference

• Added temperature detection and analysis

• Trained a classifier to detect fine-grained cooking progressions across types and temps

• Tested across lighting, oils, and pan types for generalization

• Bugbashed the feature by cooking (and eating) a lot of steak

Does Steak AI Work? (Yes.)

• Gambit estimates doneness as a % toward your target

• Spots crust formation and suggests flip timing

• Detects butter foaming and recommends basting

• Tracks temperature management, when to raise, lower

• Reads the pan environment ie. shimmering oil or smoke

• Surfaces real-time nudges like Flip, Burning, Rest etc.

Cooking steak turned out to be the perfect proving ground: simple, universal, and surprisingly complex.

Check out our latest Steak AI demo, comment with your favorite cut and butcher shop.