Startup Blogs
Cal AI's COO says its founding team was luck. He says to pay for something else.
Two of Cal AI's founders were in high school, and one founder's mom had to co-sign the bank account. The team reached a reported $50M ARR and sold to MyFitnessPal. Founders keep asking COO Jake Castillo how to build a team like it. His answer: they got extremely lucky. Here is what he says to do instead.
Founders keep sending Jake Castillo some version of the same message. On 26 September he quoted it in a post on his newsletter: "how do I find people like the group of four you had??"
The four are Cal AI's founders, and Castillo is the COO. Two of them were still in high school when they started. Henry Langmack, the CTO, was young enough that, in Castillo's words, "Henry's mom had to co-sign for the company bank account."
That team launched a photo-based calorie counter in May 2024 and took it to what Castillo calls No. 1 in Health & Fitness. By the founders' own count it reached $50 million in annual recurring revenue. Then MyFitnessPal, the category leader, bought it. The deal closed in December 2025 and was announced on 2 March 2026.
So founders want the recruiting process. Castillo's answer: "We got extremely lucky."
The chain that put him on the team
He describes how he met Zach Yadegari, the CEO, in this order:
- His cousin sent him a YouTube video about Tabs Chocolate.
- That led him to Oliver Brocato talking about internet money and his businesses, so he followed Brocato on X.
- He "just so happened to be scrolling" when Brocato tweeted that he needed a COO.
- He cold-DMed Brocato, "never expecting a response." Brocato asked to get on a call.
- Castillo started moonlighting for Brocato at Studybuddy. Yadegari was interning there.
Take out any one step and, by his account, he never meets Yadegari. "People read our origin story as a template or something we engineered," he writes, "and that was never the case."
One choice right after that chain is easy to miss. Castillo described it in an as-told-to essay for Business Insider in October 2025. He said the plan for the new intern sounded like shadowing. He didn't think that would give the kid much. So he sent Yadegari a Slack message and asked what he wanted from the internship. They started getting on calls about how to run his side projects. Yadegari left after two months. About three months later he called Castillo about a new app that was growing fast. Castillo joined as a cofounder at the start of May 2024.
The cousin's video was luck, but the Slack message to an intern whose plan was mostly shadowing was a choice. You can make that one on Monday. He told Business Insider he joined "just because I sent a Slack message to an intern." In his newsletter he puts it more broadly: working hard and being smart "just increases your surface area for luck."
The other seats came from a camp and a timeline
Yadegari and Langmack met at coding camp when they were 10, Yadegari said in a video interview. Yadegari met Blake Anderson, the fourth co-founder, on X, according to TechCrunch. Anderson, then 24, had already built consumer AI apps such as RizzGPT and Umax.
That makes a friend from camp, a stranger met on X and a colleague from a side job. Asked in the same interview whether he was intentional with his co-founder choices and whether everyone had a different role, Yadegari said "Yes" and called the mix "very complementary." Castillo calls the whole team luck, but the only chain of luck he spells out is his own.
The early months are better documented. According to CNBC, they spent $2,000 on a social media marketing test. The app made more than $28,000 in its first month and $115,000 in its second. App stores pay out on a delay, so Yadegari and Anderson covered operating and marketing costs themselves for six months until the payouts caught up.
In July 2024 Yadegari and Langmack were interviewing their first employees from a San Francisco hacker house. Two high school students were doing the interviews.
Mike Fisher, MyFitnessPal's CEO, told TechCrunch what impressed him. Cal AI's weekly stand-up was on Sunday night because the founders were still in school. The team showed up for it. "This is someone who's not doing this as a hobby," he said. His company's numbers for Cal AI are more than 15 million downloads and over $30 million in annual revenue in under two years. MyFitnessPal calls its figure annual revenue. The founders' $50 million is ARR, a run rate. The two may not measure the same thing. Terms of the deal were not disclosed.
What Castillo says to copy instead
Castillo says founders should expect to spend. "Great people are expensive, hard to find, and already have options," he writes. "You have to pay to play."
He recommends a headhunter. Their fees usually run about 20% of first-year salary, and he calls that "100% worth it". Sourcing is mind-numbing work, he says, and most founders have not hired enough people to be good at it. When a founder tells him they won't pay that, his answer is that they won't get great talent.
The headhunter only gets you a shortlist. After that you have to sell, because "the person you want does not usually need you." They have a solid job, and you are asking them to bet on a company that might not exist in a year. He says to be generous with equity, especially for someone with the intangibles you want.
When choosing a hire, Castillo looks beyond experience. He hires for availability and for whether someone can and wants to learn. His test is one question: "Is this person actually here, and can they figure out something they do not currently know?"
Experience can even work against you, he adds. An experienced hire arrives sure of how they like to work, and that way often doesn't fit the company.
Why that advice matches the luck
Put his advice next to his story and his hiring criteria describe what luck handed the team.
Castillo was available: he had just left both his consulting job and the Studybuddy role when Yadegari called. He told Business Insider it was "perfect timing." The two teenagers had school, but they held stand-ups on Sunday nights to work around it.
Willingness to learn is harder to see from outside. The clearest sign in the record is two high school students interviewing Cal AI's first hires.
His advice amounts to this: you can't wait for that combination to arrive through a cousin's YouTube link. You have to pay for it, in fees, salary and equity. Then screen for the same two traits luck gave the team for free.
Henry's mom's signature got Cal AI a bank account. It won't get you a CTO.
