February 13, 2026

How a San Francisco Chef Uses Machine Learning for Menu Design

San Francisco is famous for its farm-to-table culture. We romanticize the image of a chef walking through the Ferry Plaza Farmers Market at dawn, squeezing heirloom tomatoes, and letting “the seasons” dictate the menu.

But behind the scenes, the city’s most innovative kitchens are relying on something far more advanced than seasonal intuition: Machine Learning.

To understand how Artificial Intelligence is actually used in high-end dining, we sat down with Marcus Thorne, Executive Chef at Aura, a critically acclaimed contemporary fusion restaurant in SoMa. Chef Thorne isn’t a coder; he’s a classically trained chef who worked in traditional Michelin-starred kitchens for a decade before realizing that the future of flavor lay in data.

Here is our exclusive conversation about how he uses AI kitchen gadgets, predictive algorithms, and machine learning to design some of the most exciting dishes in San Francisco.

Food Tech SF: Chef, let’s start with the elephant in the room. There is a lot of fear that AI takes the “soul” out of cooking. How did a traditionally trained chef end up using machine learning?

Chef Marcus Thorne: I was a skeptic, 100%. In culinary school, they teach you that menu design is pure art. But when I became an Executive Chef, I realized menu design is actually 50% art and 50% terrifying financial risk. If I put a new duck dish on the menu and it doesn’t sell, I lose thousands of dollars in spoiled meat.

I started using machine learning not to replace my creativity, but as an insurance policy. I use a commercial AI recipe generator and data platform that analyzes millions of social media posts, dining reviews, and search trends across the Bay Area. It tells me what people are actually craving before they even know it. It doesn’t take the soul out of cooking; it just takes the guessing out of it.

Food Tech SF: So, the AI tells you what to cook?

Marcus: No, it gives me boundaries. It provides the canvas. For example, last autumn, our AI dashboard flagged a sudden 400% spike in local searches for “Yuzu” (a Japanese citrus) and a growing interest in “smoky flavors” among San Francisco diners.

My team and I took those data points into the test kitchen. We didn’t ask the AI to write the recipe. We used our culinary skills to create a Smoked Yuzu Scallop Crudo. It became our best-selling appetizer for the entire season. The machine identified the trend, but we provided the craft.

Food Tech SF: That sounds like a cheat code for menu design. Can AI actually help with flavor pairings?

Marcus: Absolutely. This is the coolest part of the tech. We use a molecular pairing algorithm. It looks at the chemical compounds in ingredients. Sometimes it suggests things that sound completely insane to a human brain—like pairing white chocolate with caviar, or strawberries with fermented chili paste.

But when you look at the molecular data, they share identical flavor compounds. When we test these AI-generated combinations, they often blow our minds. It pushes us out of our culinary comfort zones.

Food Tech SF: One of the biggest topics in the SF food scene right now is sustainability. Does machine learning help you run a greener kitchen?

Marcus: It is the only way to run a green kitchen. Food waste is the silent killer of restaurants. We use a predictive AI meal planning and inventory system. The AI looks at our reservation book for Tuesday night, checks the weather forecast, and analyzes historical data.

It might tell me: “It’s going to rain on Tuesday. When it rains, your soup sales go up by 30% and salad sales drop by 50%.” Because of that data, I prep fewer salads and make more broth. Since implementing this tech, we have reduced our kitchen food waste by nearly 40%. We aren’t throwing away produce anymore.

Food Tech SF: We write a lot about home kitchen tech. Can a regular home cook use these same principles?

Marcus: 100%. You don’t need enterprise restaurant software to cook smarter. The principles are the same: precision and waste reduction.

If you are cooking at home, you should be using a smart thermometer. You should be using an AI meal planning app to ensure you only buy the groceries you actually need. And if you have a bunch of random leftovers on a Thursday night, drop them into ChatGPT or an AI recipe generator and ask it to make a flavor pairing. You are essentially doing exactly what I do at the restaurant, just on a smaller scale.

Food Tech SF: You also use a lot of “Smart Appliances” in the kitchen. Are those just gadgets, or are they essential?

Marcus: They are essential for consistency. We use commercial smart ovens with internal cameras and humidity sensors. I can program the exact thermodynamic profile for a specific roast chicken into the cloud. It means that whether my lead sous-chef is working, or a new line cook is working, the oven ensures the chicken comes out at the exact same moisture level every single night.

In a city like San Francisco, diners expect perfection every time. Smart hardware combined with AI software is how we deliver that.

Food Tech SF: Finally, where is this going? What does the menu of the future look like?

Marcus: Hyper-personalization. Right now, everyone gets the same menu. In the near future, you will sit down at my restaurant, scan a QR code, and the menu will rearrange itself based on your digital profile.

If the AI knows you are on a keto diet, it will highlight the low-carb options. If it knows you love spicy food, the kitchen will get a digital alert to increase the chili levels on your specific order automatically. The future of dining isn’t robotic; it is deeply, intuitively personal.

3 Key Takeaways for the SF Foodie

AI Drives Trends: The next viral food trend you see in San Francisco was likely identified by a machine learning algorithm months before it hit the menu.
Sustainability is Data-Driven: The best way restaurants are fighting climate change and food waste is through predictive analytics that stop over-ordering.
Molecular Pairing: AI is breaking traditional cooking rules by pairing ingredients based on shared chemical compounds, creating flavors no human would have guessed.