Brian Vallelunga and the problem that survived the startup

Brian Vallelunga’s early sales budget came wrapped in foil. He took friends who ran startups to Chipotle, offered to buy whatever they wanted, and asked them to try Doppler. The running bill reached roughly $1,000. For a company concerned with keeping digital keys under control, its first distribution strategy depended on a rather less guarded proposition: lunch was on Brian.

The burritos are an appealing detail because they bring an abstract business down to table height. Before the funding announcements and the billions of secrets moving through software, there was a developer asking another developer to give something a chance. The transaction required attention, a working product, and enough hunger to accept the invitation.

The burrito budget

Doppler’s first paying customer was Tonebase. Vallelunga met its founders through mutual friends, introduced the product through those lunch conversations, and later credited their feedback with helping shape its foundations. The useful part of the meal continued after everyone had gone back to work. Someone tried the software in circumstances its maker could no longer control.

That is the detail worth keeping beside the restaurant receipt. Early customers can flatter a founder simply by showing up. They become more valuable when they encounter something awkward and explain it. For a tool intended to live inside another engineer’s routine, the quality of that conversation matters as much as the neatness of the demo.

Vallelunga had already spent years making things by then. Doppler arrived after earlier ventures, a stint at Uber, and a business idea that refused to become the business he wanted. The path has a useful untidiness. His eventual company grew out of the work required to keep another project running.

Before the keys, a flip phone

In elementary school, a friend asked why a movie could not be put on a phone. Vallelunga went home, converted a movie into the format his flip phone could play, and shared the result. It is a small incident, but a revealing one. A limitation presented as a question became an evening’s building project.

Science fairs followed. One project involved a very light craft that could fly without wings or engines. His interests already extended beyond the screen. The attraction seems to have been the opportunity to make an improbable thing work, whether the object was a phone, a flying contraption, or, later, a piece of software for other developers.

At UC Santa Cruz, where he studied computer science, entrepreneurship became a public activity. In 2017, he was among 19 people invited to a University of California lunch with venture capitalists. The participants were selected randomly from more than 340 contributors to the university’s entrepreneur campaign. Even a career retrospective needs to leave a little room for a draw.

At the time, the university described him as Miza’s co-founder and chief technology officer. Miza was a premium advertising network working to protect publisher revenue. His public project list also includes experiments in coupons, local food discovery, analytics, and delivery infrastructure. He was trying different problems on for size long before secrets management supplied a durable fit.

The project failed. The nuisance had customers.

While working at Uber, Vallelunga was also trying to build a crypto machine-learning marketplace. That project supplied a less glamorous problem than its premise: keeping track of the credentials that let its software talk to other services. A useful application needs these little permissions constantly. Managing them can consume an unreasonable amount of the builder’s attention.

On a trip to Mexico, he faced the possibility that the marketplace would not work out. Back in San Francisco, a Stripe Atlas dinner gave him a roomful of founders and developers to question. Other people recognized the difficulty. A problem that had seemed like evidence of his own poor engineering began to look like a shared frustration.

One attendee’s urgency made an impression: she wanted a solution within days after repeated outages. That demand was more useful than a general expression of interest. It attached the problem to a person whose week was already being disrupted. The marketplace could be abandoned; the work it had exposed still needed doing.

There is a temptation to tidy this into a story about perfect foresight. The sequence suggests something more practical. Vallelunga encountered a nuisance, asked whether it existed elsewhere, and discovered people willing to talk about the consequences. Recognizing the smaller problem gave him a different reason to keep building.

Make the vegetables taste like candy

Vallelunga has described his product ambition through an unusually domestic analogy: make security as appealing as developer productivity. Vegetables and candy supply the shorthand. The point is behavioral. A protective tool earns its place when the person using it feels that it helps them get through the day.

Software secrets are credentials for other software systems. An API key might let an application use a payment service; a database credential might unlock a store of customer information. The word sounds theatrical, but the daily work is largely practical: knowing where a credential belongs, who can access it, and which application needs it next.

That work is easy to underestimate because a key is small. Its permissions can be considerable. Copying credentials between people, environments, and services creates opportunities for confusion. Vallelunga’s argument puts convenience close to the center of security: the reliable process has to survive contact with the way developers actually work.

“developers need to like using it.”

Brian Vallelunga · The Secure Developer

The sentence is almost comically plain for an industry that enjoys inventing terminology. It also sets a demanding standard. A tool can have the necessary features and still be irritating enough to invite a workaround. Doppler’s product premise asks engineers to prefer the careful path because it makes their work easier.

Vallelunga also sounds less interested in performing a constant online persona than the founder stereotype might suggest. He says he enjoys talking to people, while social media holds less appeal. The distinction sits comfortably beside the early lunch strategy. A conversation gives the builder something a public announcement rarely supplies: an immediate, specific objection.

His prescription also includes some patience with the learning process. A developer’s own habits can feel universal until another team uses the product differently. Building for those differences means listening beyond the first enthusiastic response. The lunch table gets a founder into the conversation; the product has to keep earning the invitation.

From one room to billions of reads

Doppler joined Y Combinator’s Winter 2019 batch. Getting there had taken repeated attempts: Vallelunga’s fundraising account records six rejections before an acceptance on his seventh application. The number is a useful corrective to the compressed version of a founder biography, where one institutional name can conceal years of trying to get through the door.

In April 2022, Doppler announced a $20 million Series A led by CRV, with participation from existing investors including GV, Sequoia Capital, and Y Combinator. The announcement reported more than 15,000 teams using the service and over 1.5 billion secrets synchronized each month. These were dated company figures, describing the business at that stage.

By the fifth-anniversary reflection in July 2023, Vallelunga was looking back at two people working from a one-room apartment. The company had moved from serving 600 secrets a month to more than 11 billion. Its code had also moved from CoffeeScript to TypeScript, a detail he presented with a programmer’s anticipation of judgment.

The recollection lingered on people as well as throughput. Colleagues had become friends. One team member had traveled with a partner to spend a weekend with another colleague simply for the pleasure of it. For a business measured in automated operations, he chose a strikingly unautomated example of what had made the years worthwhile.

His own description was optimism, tempered by the occasional need to look backward. That is a useful pause in a career otherwise full of new projects. The apartment, the early language choice, and the customer conversations make the large numbers legible. They show what had to exist before the scale could arrive.

A different chair, another hard problem

In June 2025, Amber Britton became Doppler’s CEO and Vallelunga moved into the chairman role. His announcement credited her work building marketing and driving the company’s go-to-market efforts. The handover gives the story another named builder, with a different set of responsibilities. A founder’s original product insight is only part of what a company needs.

Vallelunga’s newer venture is Vantheon, an applied AI research lab based in Austin. Its stated subject is how artificial intelligence could change investing across economic classes. The work sits at the intersection of AI, market design, and incentive engineering, with an ambition to make tools associated with institutions accessible more broadly.

His recruiting language there is impatient with elaborate organizational machinery and emphatic about engineers owning decisions. It is a sharper register than the burrito story, but the preference for hands-on building is familiar. Vantheon’s questions return him to markets and technology after the earlier marketplace led, indirectly, into security.

There is also room for making things without a business model attached. He creates iPad art and prints it on metal. That feels consistent with the childhood phone experiment: a digital possibility becomes an object someone can actually look at. The commercial projects are larger, but the pleasure of bringing something into existence remains recognizable.

For now, the career contains both an established company under new leadership and another research effort with open questions. The failed marketplace did not provide the ending Vallelunga had planned. It left him something useful to work on. He paid attention, asked around, and eventually started buying lunch.