Our Story
Zach Suh & Sahil Jagtap · May 2026
Not a market analysis. Not a customer interview deck. A Friday night at Old Dominion Pizza in Fairfax, Virginia, watching the phone ring out four times in twenty minutes while the kitchen worked through a queue.
That is not where most companies tell you they started. It is where this one actually did.
We met three years ago at the George Mason campus incubator. Sahil was deep in AI infrastructure, shipping production code into LangChain, Supabase, and LangFuse, and had just sold his last company, Airstitch. Zach was running a meal delivery business for senior living communities, designing menus, doing pickups himself, learning where the operational seams break.
We had built three things together by then. We knew how each other worked under stress. We knew when to ship and when to throw something out. What we did not have was a problem big enough to drop out for. Both of us were on track to graduate. Both of us were doing well enough on the side that we could have kept going as we were.
What we found in the next two months changed that.
Over eight weeks we drove around Fairfax, Old Town, and Dunn Loring. We sat in pizza shops and Indian places and Thai spots between 6 and 9 PM. We ordered food. We stayed. We watched.
The pattern was the same everywhere.
The phone rings. Someone behind the counter glances at it. Nobody picks up. The cook can't, the cashier can't, the manager is checking on a delivery driver. The phone keeps ringing for twenty seconds and then it stops. Sometimes it rings again two minutes later and somebody answers it sounding distracted because they are holding a tray of curry in the other hand and trying to read the back of a takeout box at the same time.
When we asked, the answer was always some version of the same sentence: “If we picked up every call, we would have to stop cooking.”
We did this in 20+ restaurants. We were not pitching anything. We were not asking what they thought of voice AI. We were asking how their phones worked, what was breaking, and where the money was going.
Old Dominion Pizza is a family-run shop on Main Street in Fairfax. The first time Zach walked in, the owner sat down with us at the back booth for ninety minutes.
He told us 97 out of every 100 of his orders come through the phone. He told us he shuts down Uber Eats mid-shift on Fridays because the kitchen physically cannot handle one more order. He told us he loses orders every single week because the phone rings and nobody can answer it.
He did not want a chatbot. He did not want a phone tree. He did not want a call center service. He wanted his phone to do the thing it was supposed to do: ring, get answered, take the order right, drop it into the kitchen.
That conversation reframed the whole thing for us. This was not a voice AI product. It was direct revenue infrastructure for a business that runs on margins of 5 to 15 percent. The phone is where their highest-margin demand is showing up. It just needs to actually work.
We started writing code the next week.
We have heard this question a few times now. Why us.
We are not the most experienced restaurant operators in the world. There are people who have run restaurant groups for thirty years. We are not the most experienced AI engineers either. There are teams at OpenAI and Google with more engineers than our entire pilot list has restaurants.
What we have is a specific overlap that almost no team building in this space has: an operator who has spent the last year inside food businesses, paired with an engineer who has spent the last year inside production AI infrastructure, both willing to drop out for it.
Zach. suhzachary.com Zach grew up cooking for his grandmother through her late-stage Alzheimer's. The kind of cooking where you learn what someone can taste and what they can't, what makes them eat and what makes them push the plate away. Food was how he kept her present.
He turned that into a private chef service for seniors, then HomePlate, a meal subscription business that turned senior living facility kitchens into delivery operations. HomePlate did $4,000 in its first month, 42 meal boxes delivered, 73 preorders, $11,000 in non-dilutive grants, and was covered by Fox News, Yahoo News, WTOP, DC News, and PatchNews. He won first place at George Mason's Patriot Pitch competition and was on a full academic scholarship.
While he was working at Browser Buddy (YC W24) in San Francisco, he hosted private dinners for YC founders and venture investors. The kind of person who walks into a room and people remember him a year later.
He is the reason every restaurant we have talked to has agreed to talk. He walks in, he sits down, he asks the right questions, and he comes back the next week with a plan. He does not send cold emails. He has not run a single ad. Every pilot we have closed was closed in person.
He is 20. He is dropping out of George Mason to do this full time.
Sahil. jagtap.tech Sahil built and sold a company before most engineers ship their first production feature. Airstitch, the world's first text-native AI chief of staff, hit 200+ active users in seven days from launch and was acquired within three months. AgentBudget, his open-source AI agent resource limiter, has 100,000+ installs with no marketing spend and a published whitepaper.
He is a Cerebras Engineering Fellow. He has shipped production code into Scribe, LangChain, Supabase, and LangFuse, four of the most heavily used pieces of AI infrastructure in the industry. He has won more than 30 hackathons, including ones at MIT, Princeton, and UPenn. He led NSF-backed wearables research with 500+ participants and 94 percent ML accuracy. He is part of the Founders Inc Artifact 2026 cohort and got into YC Startup School 2026.
He is 22, in computer science at George Mason, and is also leaving school to do this.
The two of us. We have been working together for three years. We have built and shipped multiple things. There is no overlap and no gap between what we cover. Zach is in the restaurants. Sahil is in the system. Every owner we have sold to has both of our cell numbers. We answer.
We are not the people who have spent a career in this. We are the people willing to sit in a pizza shop at 7 PM on a Friday for two months before writing any code. As far as we can tell, that is the part that determines whether this product actually works.
Old Dominion Pizza is live on Cardamom. Bollywood Bistro and Saravana Palace are next, both within the next two weeks.
We are not chasing scale yet. We are chasing one number: does a restaurant on Cardamom make measurably more money this month than they did the month before, with the same dining room, the same staff, the same menu.
Once we have that proof at three restaurants, we go to ten. Once it holds at ten, we go everywhere takeout-heavy restaurants exist. We are not in a hurry to be everywhere. We are in a hurry to be right.