GARBO / TRUST & SAFETY
Helping people make more informed decisions.
Keeping the "yikes" out of our swipes
ROLE
Founding and sole product designer responsible for shaping Garbo's 0→1 consumer background-check web app, establishing its information architecture, core search and results experiences, and design integrations with Tinder and Match to make access to public record information more approachable and actionable for online daters.
SCOPE
0→1 Consumer Background-Check Platform, B2B Dating App Integrations
STATUS
Launched and operated for ~1 year, then gracefully retired due to scaling challenges with third-party public record data and social stigma.
FOCUS
Consumer Product Design, Trust & Safety, Information Architecture, UX/UI, Front End Engineering
ORGANIZATION
Garbo.io
02 / THE BRIEF
Taking the sweat out the swipe.
When online dating became the default way to meet people, a strange kind of collective amnesia took over. In the sheer excitement of swiping and meeting strangers, the old rules of "stranger danger" were completely forgotten. We all collectively decided to overlook the risks just for the thrill of the connection. While finding your person had never been easier, keeping yourself safe felt like a parental chore—usually reduced to your mom nagging you about your whereabouts or digging a crusty canister of pepper spray out of the bottom of your bag.
The reality, of course, was much darker. Bad actors were actively weaponizing dating apps, and most people didn't know how to spot a red flag until they were already standing in the wreckage. After experiencing this firsthand, Garbo's founder decided to do something radical by building a background check that would be as fun as a swipe, clean up the apps, and finally take out the trash, hence the name.
One of the things our fearless founder hated was traditional consumer background checks. These are predatory scams are basically digital shakedowns designed to profit off your fear, operate out of murky offshore headquarters, and as a bonus, they immediately sell your personal info to an army of robocalling spammers who will never, ever stop calling your phone.
Garbo aimed to flip the script entirely, partnering directly with the National Domestic Violence Hotline and high-profile prison reformers to build a north star that would make an angel blush. As the founding and sole product designer, I stepped in to take a notoriously bleak corner of the internet into something like unearthing a secret level in a video game you didn't even know you were allowed to play.
THE CENTRAL DESIGN TENSION
How do you take a topic that’s usually a legal minefield, a privacy nightmare, and a PR hazard, and turn it into something weirdly fun to use?
03 / DISCOVERY
Clearing the landmines out of the 12-step maze.
To figure out how to make background checks feel less like stumbling blind through a Halloween cornfield with a flashlight running on dying batteries, I had to put myself in the user's shoes and actually run through the gauntlet myself. Buckle up.
First, you plunge headfirst into a grueling, 12-step digital obstacle course that weaponize the sunk-cost fallacy. The platform drags you through a punishing, soul-crushing trail of friction and every intrusive question under the sun before slamming into a paywall complete with a ticking countdown clock holding all your hard-earned data hostage until you cough up cash right on the spot.
But wait, there's more! Just when you think you’ve limped across the finish line with only a few scrapes and bruises, you’re hurled into a second guided maze stretching the length of a Saturday afternoon at IKEA. This boss level demands data you could never, ever expect to casually pull from a Tinder bio. A full social security number, home address, hometown, parents' names, and place of work, just to name a few. And it wasn't a matter of giving what you had. Every single field was a hard blocker. No SSN for some guy named Snake in a dimly lit back alley? No safety for you.
And for all that friction, the payoff was a bit like downloading Eminem off of Limewire in 2004. Matches lacked photos, used dirty data in dense, intimidating legal jargon, and worse still, they treated every minor infraction like a parking ticket as if it were a flashing neon red flag, which research showed were systemic artifacts of discrimination rather than indicators of future violence. Whatever accurate data they actually managed to salvage were sensitive goldmines gift-wrapped like a starter kit for stalkers.
The Dusty Archives (out of scope)
Trapdoor, sunk cost onboarding
Predatory paywalls
Weaponized parking tickets and minor traffic violations
Creepy, stalker-ey PII like SSN, home address, and job history
Dense, unreadable legalese
Leaves you needing immediate therapy
The Safety Filter (in scope)
Non-manipulative onboarding
Freemium safety or suggested donation without the hostage situation
Laser focused on actual threats approved by prison reformers and people who are generally outraged by everything
Actionable, transparent insights without handing a stalker an instruction manual
Simple enough for a third grader, who are smarter than us—let's be real
A post-search guide that feels like a reassuring chat with a gentle neighbor from a 90's sitcom
We had to burn the old playbook to the ground. Instead of hoarding every dusty public record on the books, I drew a hard line between old-school, punitive record-keeping and forward-looking, high-stakes prevention. If the data didn't fit that criteria, it didn't even make it to the cutting-room floor.
Partnering with the country's most accurate public record database typically reserved for corporate background checks (and loved what we were doing), they revealed a shortcut. We didn't need a federal dossier at the gate. All we needed was a phone number and a zodiac sign to lock down the birth date—low-stakes details you casually drop in the DMs before your first coffee date. Better yet, instead of drowning the user in a messy sea of false positives, the architecture cut straight through the noise to pinpoint a single, accurate match.
Image is for illustrative purposes only.
I mapped out and prototyped the entire end-to-end journey starting from the initial entry point right inside a dating app, rolling through frictionless onboarding and smart fuzzy search, all the way to delivering the search results and guiding users on what to do with that information after the fact. Before writing a single line of production code, we put this comprehensive prototype through the wringer with a rigorous two-tier research sprint. The first was deep, moderated round with four participants, followed by an unmoderated round with 30 users to stress-test every edge case. We wanted to guarantee the experience felt as crystal-clear and empowering as it was safe. This was the rock-solid foundation we need before adding a bit of unexpected magic.
04 / THE CHALLENGE
Just dropping a search bar into a dating app wasn't going to cut it.
The stress-tests uncovered three glaring barriers that went way beyond standard UI friction.
The privacy stigma was real. Users were paranoid that someone on the other end of the server was sitting in a dark room with a live microphone and will broadcast of their full search history, so the last thing they wanted to do was blindly hand over their traceable legal name and email. Filling out a basic account creation form felt like a high-stakes interrogation as if they were an underage drinker with a terribly laminated fake ID.
Instead of feeling reassured by our educational onboarding, there was a nervous energy about seeing something they could never, ever unsee. Yet, trapped right alongside that paralyzing dread was an irresistible pull of morbid curiosity. Once the paranoid users finally laid eyes on the raw results, their cognitive dissonance went off the rails as if they had wandered into a situation so unsettling it had to be a fake cardboard cutout of reality and waited for Ashton Kutcher to leap out of a fake plant, praying the whole thing was just a very elaborate joke.
Turns out, you can't just dump a raw criminal record on someone's lap and tell them to "have a nice day." To dismantle that stigma, the interface had to actively restructure how our users processed their fear.
By anchoring the design in progressive disclosure and contextual translation, I had to meet users where their panic lived. Instead of forcing them to decode clinical legal jargon alone, we broke down the cognitive load by explaining what those flags could and more importantly couldn't prove, replacing a late-night doomscroll with clear, actionable next steps beyond the boilerplate crisis language.
THE REAL DESIGN CHALLENGE
How do you build trust when they're convinced the system has a megaphone and will broadcast their search history to the whole world?
05 / THE DESIGN
From yikes to yaasss.
Armed with insights from our first round of testing, I went back to the drawing board to bridge the gap between panic and peace of mind. I overhauled the entire user journey to build real search comprehension, squash skepticism, and trade clinical warnings for genuinely grounding advice.
Instead of feeling like suspects under questioning, users were given clear context on why certain clues mattered so they walked away feeling genuinely in-the-know. I baked transparency directly into the UI by introducing confidence signals and system grading, giving people the tangible proof they needed to trust what they were looking at.
Next, we brought in an editorial team to build localized safety pathways that met users right where they were. We ditched stiff, clinical crisis language in favor of a curated resource hub with the vibrant energy of Refinery29 and Vice, transforming a sterile background check into a space where users felt genuinely seen, heard, and supported.
We put this V1.2 evolution through its paces with an additional set of 30 unmoderated participants to stress-test the experience. The results spoke for themselves.
40% of users who started out deeply skeptical about running a date check completely changed their mind after clicking through the prototype.
The remaining groups fell into two camps. They either would never, under any circumstances, want a background check involved in their love life, or survivors who had a fierce, visceral anger that we were filtering out minor traffic violations. This was fine because Garbo wasn't trying to convert the whole dating pool, just establish a safe, ethical standard for the people who didn't know they wanted that layer of protection, and were okay with our transparent approach.
With the core experience thoroughly validated, the tightest turnaround imaginable, and only one engineer to help, I rolled up my sleeves to write the entire front-end codebase myself (backed by only one year of Codecademy) with zero oversight, and got the platform and high-stakes Tinder integration all the way across the finish line. It was like letting a toddler with sticky fingers deploy production infrastructure of a major corporate safety net while Tinder's engineers watched in quiet horror. I still can't believe I did that.
06 / THE PIVOT
Is this the real life? Is this just fantasy? (Caught in a landslide)
With the fully coded platform completely built, baked, and ready to launch, we encountered a minor out-of-body experience. Giddy with excitement, Tinder executives did what anyone would do and searched their own names. The system promptly matched their clean records to several sex offenders with zero identifying details other than a matching eye color. Not your average Tuesday. On top of it, "The Tinder Swindler" documentary had just come out on Netflix and they were not in the best of moods.
Image is for illustrative purposes only.
Without the luxury of time for finger-pointing, it made total sense for us to completely blow up the idea of a single confident match. Instead, we let human intuition take the wheel, letting the system throws its metaphorical hands up to say, "Honestly? It's one of these people. You decide." And I promptly redesigned the results flow to match this new direction.
We traded a dangerously cocky algorithm for a humble, multi-match framework, saving the launch at the eleventh hour, and proving that the best safety design is the one that trusts the user to connect the dots.
WHAT CHANGED
Before
After
Single identity match
Multiple possible matches
Trusting the user to connect the dots
07 / THE LAUNCH
Garbo in, Garbo out.
Following exhaustive pre-launch testing with our waitlist, we officially unleashed the mobile-first platform into the wild and steered the design as the product evolved post-launch, which by the way, turned into a full-blown media phenomenon. Paraphrasing headlines such as "Tinder has a background check now and you can search anyone you want," sparked conversations across The New York Times, The Wall Street Journal, TechCrunch, Fast Company, and my personal favorite—Reddit. Users rushed through the doors, driving a massive 5% landing-page conversion rate in week one alone.
Adoption snowballed across other dating apps like Bumbl and Grindr, before revealing an urgent, unmet need in peer-to-peer digital marketplaces like Facebook Marketplace and Craigslist. Naturally, as we scaled, the data architecture grew exponentially more complex.
While Garbo ultimately retired after a year, we proved something undeniable. Platforms come and go, but we pioneered a non-negotiable blueprint for the future of online trust and safety. Once consumers got a taste, it was all they ever wanted. Even after winning over our fiercest critics, the social era wasn't quite ready for a glow-up this ambitious. While our engine was robust enough to clear you for a sterile corporate cubicle at Bank of America, it collided with the messy, human reality of an instant match when we rely on the accuracy of police data entry where someone's greasy donut hands accidentally slip on the keyboard and type a name into the address field, or misfile a misdemeanor because typing is hard.
Garbo's impact echoed across the industry., and long after our servers went dark, the legacy remains. We set a bold, human-first precedent for keeping the "yikes" out of our swipes.