Work

I don't have 16 years of experience in product. Or engineering. Or sales. But I've spent 16 years doing all of it at once.

Started at 17 flipping burgers and knocking on doors. Got into tech through freelance translation, then joined a four-person software company and stayed for thirteen years - built nine products, killed a few, grew it from four people to thirty, and ended up running the place. $10M+ ARR on less than $20K/month in infrastructure. No outside funding.

Now I'm building Junto from scratch. AI recruiting platform. Same instincts, different bet.

16 years · 9 products · still figuring it out

2025–Present

Junto

Co-Founder

After 13 years of building someone else's company, I wanted to build my own.

Junto is an AI-native recruiting platform. 12 integrated modules built from scratch. The core differentiator is a patent-pending system for near-real-time deepfake detection in live video interviews, combining adaptive biometrics, multimodal analysis, and tamper-proof audit logging. I designed the system architecture, wrote the patent application, and built the first working prototype myself.

€800K+ in public R&D grants secured. Partnerships with universities and compliance labs. Building for the EU AI Act from day one, because regulation is coming whether the industry is ready or not, and the companies that treat compliance as a feature will win.

I'm doing everything again. Product, engineering, fundraising, hiring, ops. The difference is this time it's intentional. Thirteen years of running a portfolio taught me what kind of company I actually want to build: small team, deep technology, real revenue, no fluff.

2012–2025

Vitzo

2022–2025 CEO

2018–2022 Head of Product

2014–2018 Product Manager

2012–2014 Founder's Associate / First Non-Developer Hire

Simple video downloader for mass adoption.

150K DAU · 50K daily partner installs

This is where my career really started. With toolbar offers.

When I joined, both VDownloader and Viddly (then called YouTubeDownloader) made money the way everyone did in that era: bundled installer offers. Think Ask Toolbar and browser hijackers. It worked. Until it didn't. Revenue per install was declining quarter over quarter, and the entire model depended on users not paying attention. That's not a business. That's a tax on confusion.

I built the case for subscriptions internally, then did the work: payment infrastructure, pricing tiers, trial mechanics, a new website, and a GTM plan. All while finishing my CS degree. We A/B tested every step of the conversion funnel, from install screen to first payment prompt. The transition wasn't clean, but within two quarters we'd replaced the toolbar revenue entirely with recurring subscriptions.

Distribution partnerships

Growth came from an unusual place. We partnered with companies like GOM Player to cross-promote installs. They wanted Latin America, where Viddly had a massive user base. We wanted APAC. No money changed hands. Just installs for installs. At its peak we were trading 50,000+ daily installs and Viddly had over 150K daily active users.

We also built automated engagement loops that re-activated dormant users and fed them back into the conversion funnel. Distribution doesn't always cost money. Sometimes it costs inventory. Finding the right trade is a product problem, not a marketing one.

Viddly was positioned as the affordable, simple option. Fewer features, lower price, broader audience. It complemented VDownloader rather than cannibalizing it. We eventually renamed it from YouTubeDownloader (YouTube didn't appreciate that). Between them, Viddly and VDownloader have been installed over 50 million times.

Full-featured video downloader for power users.

50M+ combined installs with Viddly

The company's original product. Already established when I joined, with over 300 supported video sites, every format option you could ask for, and a power-user base that cared about things like bitrate selection and batch queuing.

My job was to keep it healthy while everything else grew around it. Coordinating releases with engineering, managing the backlog, running retention experiments to understand what kept power users paying year after year. No single dramatic inflection point. Just years of steady product work on something that reliably generated revenue and gave us the margin to invest in newer products.

VDownloader was the premium tier. Viddly was mass-market. Running both taught me how to serve different segments from the same codebase without diluting either one. Same technology, different packaging, different price, different users. That's portfolio management at its most basic.

Safe, ad-free video sharing for classrooms.

$0 → $100K MRR in 3 months · B2B pivot to mid-six figures

This was my biggest bet.

SafeShare strips ads, comments, and distractions from YouTube and Vimeo videos, making them safe for classroom use. When we acquired it, over 100,000 teachers used it every day. It also made exactly zero dollars.

Turning it into a business

The founder was skeptical about charging. He worried a paywall would drive educators away. I looked at the data differently: 100K daily actives with no revenue meant we had product-market fit and no business model. That's a solvable problem.

We spent six months rebuilding the platform. New subscription system, new onboarding, new everything. Pure SaaS. I designed the trial flow, set up the conversion funnel, and iterated on pricing until the numbers worked.

Three months after launch: $100K MRR.

The pricing change was the visible move. The real work was everything underneath. We ran multivariate tests on the onboarding flow, experimented with trial lengths and feature gating, and built retention loops that tied product engagement directly to expansion revenue. The product didn't just get a price tag. It got a growth model.

Scaling through B2B

We noticed clusters of users from the same school districts. Teachers were adopting individually, but the usage patterns screamed institutional. I ran a cohort analysis that confirmed it: district-level adoption was 4x stickier than individual signups.

We pivoted to a seat-based model for schools. Tiered pricing, admin tools, bulk management. This was months before COVID made e-learning everyone's priority.

Revenue jumped from low six-figure MRR to mid-six figures. Today, SafeShare has over 1.3 million registered accounts and more than 45 million SafeViews created worldwide.

Clipboard manager with built-in editing tools.

1M+ users incl. US government agencies

Started as an internal tool during the period when we were expanding the portfolio from two products to four. We needed a better clipboard manager, evaluated what was on the market, found nothing worth using. So we built one for ourselves, validated that the gap was real, and shipped it publicly.

ClipClip now has over a million users, including US government agencies. For a product that started as a side-of-desk build, the signal was clear: sometimes the best product ideas come from scratching your own itch and discovering the itch is universal.

Clean, safe video sharing for a general audience.

+40% market share via acquisition

Acquired as SafeYouTube.net, a direct SafeShare competitor. The thesis was simple: buying the number-two player is cheaper than outspending them. The deal increased our market share by 40%.

We rebranded it to VideoLink (YouTube trademark restrictions) and repositioned it for a general audience. Same core technology as SafeShare, without the school-specific focus. One codebase, two brands, two markets.

Screen capture with built-in editing.

Sunsetted

Products we built and intentionally killed or consolidated. Showing these because knowing when to stop is part of the job.

Online YouTube downloader.

Built it. Ran it. Killed it.

YouTube kept advancing their streaming infrastructure in ways that made online downloading progressively harder. Anything above 720p required serious backend work, and the costs scaled in the wrong direction.

The alternative was what every other free downloading site did: plaster the page with 20 ads and hope the margins work. We ran the unit economics three ways and none of them closed without destroying the user experience. If the only path to profitability requires making the product miserable to use, the product is already dead.

We reallocated the team toward SafeShare, which was showing early traction at exactly the right time. That reallocation funded the SaaS rebuild that took SafeShare from zero to $100K MRR.

Native mobile app for SafeShare.

Built it. Learned the hard way.

The logic seemed sound: SafeShare had massive daily usage, so a mobile app should capture that audience on the go. What we missed was the environment. Schools run on Chromebooks, not phones. IT administrators had zero appetite for whitelisting a third-party mobile app. And teachers weren't watching videos on their commute. They were projecting them in classrooms.

We caught it early. Usage data told the story within six weeks. We killed it before it became a money pit and doubled down on the web platform, which is where every dollar of SafeShare revenue eventually came from.

Chrome extension for visual annotation and collaboration.

Not a failure. A cleanup.

ScreenLink was a Chrome extension. ScreenClip was a web platform. Both did overlapping things under separate brands. Users were confused about which one to use. The team was splitting effort across two codebases that solved the same problem.

In early 2023 we unified them under the ScreenClip name. Support tickets about "which tool do I use?" dropped to zero. Sometimes the right product decision is removing a product.

Loop any YouTube video, endlessly.

20K daily active users for a product that did exactly one thing: loop a YouTube video.

We tracked platform risk as a core metric. Every quarter I'd audit which product features overlapped with roadmap signals from YouTube, Google, and the major browsers. EndlessVideo was always in the red zone. The entire value proposition was one checkbox away from extinction.

So we planned for it. Migrated the highest-value users toward our other tools. Built re-engagement flows that cross-sold into Viddly and SafeShare. By the time YouTube added native looping to their player, we'd already moved 60% of EndlessVideo's active base into other products in the portfolio.

The shutdown was a non-event. That was the point.

2011–2012

Upwork

Freelance Translator

After that I discovered you could get paid to work from your bedroom. Freelance translation on Upwork, English to Portuguese and back. Medical device manuals, toy labels, audio transcriptions, whatever came in. I used the money to fund my education and never went back to working on-site. When I joined Vitzo, it felt like a natural fit, four people working remotely, building software. No office, no commute, just the work.

2010–2011

Portugal Telecom

Door-to-Door Sales

Then I spent eight months knocking on doors for Portugal Telecom, which is exactly as glamorous as it sounds. You learn a lot about persuasion when most people close the door before you finish your sentence. It was rejection on repeat, every single day. Turns out that's decent training for product work, where most of your ideas also get rejected, just more politely.

2009–2010

McDonald's

Restaurant Worker

Before any of this, there were burgers. I worked at McDonald's part-time during high school, the kind of job where you learn that customer service is mostly about keeping your composure when someone wants a Big Mac with no bun. Ten months of that and I had a solid work ethic and a permanent aversion to deep fryers.