How We Got Into EASY Residency

If you work in crypto, you’ve probably heard of YZi Labs.

Its EASY Residency is one of the most sought-after incubators in crypto. We were told that only around 3–4% of applicants get in.

It took us two tries. Looking back, I believe the difference was learning to deliver the outcome our customers actually needed: finding serious vulnerabilities before someone exploited them. We did that for YZi-backed teams, who validated our findings and paid us for the work. By our next interview, we had evidence that we could solve a real customer problem.

Here’s how that changed both our application and the way we built ZeroDrift.

Before the application

ZeroDrift started with a simple question in early 2026: as AI gets smarter, can it find security vulnerabilities as well as humans, or even better?

Chris, my cofounder, had just left a leading blockchain security firm. I was about to leave another.

Security expertise was scarce and expensive. We wanted AI to make it as readily available as electricity, so teams could afford to review every code change.

Our MVP was a GitHub bot that audited pull requests and automated security tests in CI/CD. Once we had a working demo, we tested it with pilot customers. It found critical bugs that leading security firms had missed.

That felt like enough of a starting point to explore fundraising and incubators. So I researched the options and applied to every program that seemed like a reasonable fit.

Rejection season

We started applying in February. There were endless forms, dozens of deck revisions, and investor meetings at all hours.

Most applications and investor conversations ended the same way: rejection.

We documented everything: our applications, decks, traction, meeting transcripts, and the questions I struggled to answer. Then we used those notes to prepare for the next conversation.

Each interview taught us something. Sometimes it exposed a gap in the business; sometimes it showed us that we were explaining ZeroDrift badly.

Fundraising takes practice.

At the end of February, we found our first investor: Christian. After so many rejections, having someone believe in us felt incredible.

Christian introduced us to Ricky, who would later lead our investment at YZi Labs.

Applying to EASY Residency

After a quick call, Ricky encouraged us to apply. Applications for Season 3 were still open.

We ran more evaluations of our security bot on EVMbench and went through the deck several more times.

The application itself was straightforward. I gave my AI agent access to our existing materials and had it fill out the form in Chrome through a browser connector. Then I checked the completed application.

Over the weekend, we ran mock interviews with Liz, our earliest supporter. We recorded each session and had our deck-coaching agent analyze the recordings to help us improve.

The first interview

On the evening of March 17, we had our first interview with four or five investors from YZi Labs.

We spent the first ten minutes walking through our deck. Then came the questions.

They were mostly the ones you would expect if you had done your homework:

• Do you have a working demo?

• What makes you different from your competitors?

• What is your business model?

• What traction supports your assumptions?

The challenge was answering concisely and convincingly, with evidence to back us up.

After another ten minutes, the questions were over. The investors thanked us, and the meeting ended.

Months of preparation. A twenty-minute interview.

About a week later, an email arrived.

read image descriptionALT

We had been rejected 😢

Reading it hurt. Sharing it with the team was even harder.

I asked Ricky for feedback. He explained that although the idea was innovative, we needed stronger traction to validate the product.

Getting into EASY was clearly not going to be easy.

But we were determined to find out whether our idea worked, regardless of the application outcome. So we kept building.

Deliver outcomes first

By Q2 2026, AI auditors were everywhere: Zellic’s V12, Cantina’s Apex, CertiK’s AI Auditor, and a growing list of other tools. Even for us, it was hard to keep up.

We needed to give teams a concrete reason to trust ours.

We began to question our approach: keep developing the product, market it, find more customers, and iterate. How long would it take to prove that people valued what we were building?

We went back to a simpler question: beneath all the workflows, dashboards, guidelines, and best practices, what did our customers actually want?

In security, a concrete answer was finding a serious vulnerability before someone exploited it.

We could use the tools we already had to deliver that outcome immediately.

Prove it in live systems

Choosing where to look mattered as much as the search itself.

We wanted projects that were actively shipping, had substantial usage or trading volume, and ran complex products where a vulnerability could have a meaningful impact.

We started with projects backed by YZi Labs. We knew that useful findings in companies the investors already understood would be easier to evaluate.

We filtered out inactive projects and infrastructure outside our scope, then prioritized the remaining candidates by potential impact.

Then we pulled their verified source code, built the context our agents needed, and put our AI auditors to work.

Over the following weekends, we burned through tokens at an alarming rate. One Max plan after another hit its limit.

The work paid off. Our AI auditors found high-impact vulnerabilities that previous audits had missed in live systems across Sign, Renaiss, Flap, and other platforms.

Ricky introduced us to the teams, who validated our findings, paid bounties, and explored further collaboration. We now had concrete results to show investors.

At Sign Protocol, for example, we found that one signed authorization could be reused to create attestation records beyond what the signer intended. Services relying on those records for identity checks or access permissions could end up trusting records that should never have existed. Sign implemented fixes in its codebase and continued working with us to verify its live deployments.

EASY’s network across the BNB Chain ecosystem could help us bring that work to more teams.

The final rounds

In June 2026, several months after our first interview, we heard from YZi Labs again. We were invited to interview for Season 4.

By then, we had just reached break-even. We thought this would probably be our last incubator application, whatever the outcome. The word “accelerator” finally made sense to me: we had built momentum of our own, and the right support could help us move faster.

We had a casual conversation with Ricky, followed by two more interviews.

The first, on June 11, felt much easier than our March interview. We had been sharing regular progress updates in our group chat with YZi Labs, so the investors were already familiar with what we had been doing.

We walked them through our findings, how we approached customers, and what we planned to do next. This time, we had concrete examples to discuss.

The second interview, on June 17, was led by Ella, the Managing Partner of YZi Labs.

She asked more than ten questions covering legal matters, technology, customer onboarding, business models, and our competitive advantages.

I was nervous, and I did not have a perfect answer to every question. But I kept reminding myself to be honest: admit what I did not know, and explain where I saw things differently.

The conversation pushed me to think more carefully about the business. It felt like a coaching session: direct questions, challenged assumptions, and very little room for vague answers.

Two weeks later, around midnight on a Sunday, Ricky messaged us.

Five months after our first code commit, we had been accepted into EASY Residency 😊

I barely slept that night. I called a few friends, talking too quickly, my voice still shaking a little. Then I hugged my girlfriend so tightly we could barely breathe.

That’s my advice to other founders: keep building, even when you get rejected. Use the skills you already have to solve one concrete problem for a customer. Learn from the feedback, and let the results guide what you build next.

You don’t need a 15-person team or a fancy website to launch. We started ZeroDrift as three twenty-somethings bold enough to try before we had it all figured out.