Two brothers in Stockholm almost passed on a great hire because his resume looked ordinary. They talked to him anyway, and inside a few minutes the grit was obvious. That near-miss is now a company with four million dollars in the bank.
Fika Jobs announced its $4 million pre-seed on Tuesday, June 23, and the pitch is blunt: the resume is a lousy way to find out who somebody actually is. The platform pairs AI interview agents with short-form video profiles, landing somewhere between LinkedIn and TikTok. Instead of firing applications into a void, you build one living profile, sit through a roughly ten-minute video interview with an AI agent, and let employers come find you. The agent runs on Google's Gemini models. Your answers get sliced into short clips and organized into something a hiring manager can actually watch.
Here's the part that makes it interesting. Most of the AI hiring crowd is building tools for the people doing the hiring. Fika flipped the camera around.
The black box everyone hates just got worse
You know the drill. Hours spent tailoring a cover letter nobody reads. An application that vanishes into an applicant tracking system. Maybe a rejection email six weeks later, maybe nothing at all. That experience was bad before generative AI showed up. Now it's worse, because employers are drowning in machine-written applications and leaning on machine screeners to cope. Bots talking to bots, with a human somewhere in the middle losing the thread.
Fika's co-founders, brothers Jakob Dubois and Alexander Dubois, hit this wall while running their previous startup, a social app called Gaff. The candidate they nearly skipped became the proof of concept for a different idea. Some of the traits employers say they care about most, drive, curiosity, the way someone thinks out loud, just don't survive contact with a PDF.
So they built the opposite of a PDF. You connect your LinkedIn, the AI reads your background and writes questions tuned to you, and then you talk. Not to a recruiter who's seen forty candidates that day and is running on cold coffee. To an agent that asks the same thoughtful questions every time and never gets bored by candidate number forty-one.
That consistency cuts two ways, and it's worth sitting with. A tireless interviewer treats the first candidate and the last one the same, which is more than most exhausted hiring panels can claim. It also means whatever the agent rewards, it rewards every single time, at scale, without a bad mood or a good one to break the pattern. Design the questions well and you get fairness. Design them badly and you get bias on autopilot.
Why Fika points the AI at you, not your boss
This is the real wager. Companies like Mercor, Maki, and Alex are racing to help employers source, screen, and rank candidates faster. Useful work. Also a crowded room, with deep-pocketed incumbents and a dozen seed-stage clones all selling the same promise to the same overwhelmed recruiting teams. Fika is betting that the bigger unlock sits on the other side of the table, where almost nobody is building.
Picture a pool of people who've already been interviewed and assessed, sitting in profiles employers can browse and revisit whenever a role opens up. You interview once. Your profile keeps working while you sleep. A hiring manager scrolling that pool gets a read on communication and presence in the first thirty seconds, which is exactly the stuff a resume hides.
The model borrows its logic from how dating apps and creator platforms already work. You don't reapply to be visible. You maintain a profile, and the system surfaces you when there's a match. Applied to hiring, that flips the entire burden. Today the candidate does the chasing, over and over, role by role. Fika wants employers to do the discovering instead, pulling from a standing pool rather than waiting for inbound applications to pile up.
If it works, the people who benefit most are the ones the old system fails hardest. Early-career candidates with thin resumes. Career-switchers whose experience doesn't map cleanly to a job title. Anyone whose strength is how they show up in a room rather than how they format a bullet point. That's a real population, and nobody's serving it well right now.
The risk is just as real. Video interviews can smuggle in bias the same way humans do, rewarding polish, confidence, and a certain accent over substance. There's a reason regulators in New York and the EU have started circling automated hiring tools with audit requirements and transparency rules. Fika will have to prove its agent measures what it claims to measure, and prove it to skeptics who've seen plenty of hiring AI promise objectivity and quietly deliver the opposite. Personality over performance. Signal over showmanship. Easy to say on a pitch deck, hard to ship.
The math behind a $4M bet on a profile you build once
Four million at pre-seed is a meaningful check for a Nordic startup in 2026, especially one going after a category as picked-over as hiring. The money goes toward building out the platform, growing the team, and prepping for a wider launch later this year. Modest goals, stated plainly, which is refreshing in a space full of companies promising to fix work forever.
Stack it against what else is moving in the Nordic early-stage market and the number reads as a healthy bet rather than a moonshot.
Detail | Fika Jobs |
|---|---|
Round | Pre-seed |
Amount | $4 million |
Announced | June 23, 2026 |
HQ | Stockholm, Sweden |
Founders | Jakob Dubois (CEO), Alexander Dubois (CTO) |
AI engine | Google Gemini models |
Use of funds | Platform, hiring, wider launch in 2026 |
Compare that to the new pre-seed funds writing these checks across the region. The under-25 crowd at Wave Ventures just closed €10 million for exactly this stage, and community-driven firms like byFounders are flush with fresh capital for the New Nordics. There's money looking for founders who can build something people use daily. Hiring qualifies. Everyone job-hunts eventually.
The harder question is revenue. A two-sided marketplace is brutal to bootstrap, because neither side shows up until the other one's already there. Candidates won't build profiles for an empty employer pool, and employers won't pay to browse an empty candidate pool. Fika has to solve that chicken-and-egg problem with $4 million and a wider launch a few months out. That's a tight window for a cold-start marketplace, and it's where most platforms in this category quietly stall.
Sweden keeps minting consumer-shaped software
Notice the shape of this thing. It's a consumer product wearing enterprise clothes. The candidate experience feels like an app you'd open on the couch, swipeable, video-native, fast. The buyer is a company. That blend, consumer polish aimed at a business workflow, is a Stockholm specialty. Spotify did it to music. Klarna did it to checkout. The instinct to make boring software feel good runs deep in this city, and it tends to travel well beyond it.
Fika is doing it to the job hunt, which has resisted every attempt at reinvention for two decades. Job boards got prettier. The underlying misery stayed the same. A short-form video layer with an AI interviewer underneath might finally change the texture of applying for work, or it might be one more clever wrapper on a broken process. Too early to call.
What's not too early to notice is the founders' read on AI's role. They're not using it to filter you out faster. They're using it to give you a fair shot at being seen. In a year where most AI hiring news is about automation replacing judgment, that framing stands out, and it's a sharper wedge than the team is getting credit for. The product is a marketplace. The story is about dignity in a process that strips it away.
Whether anyone actually wants to be interviewed by a robot
The honest open question is behavioral, not technical. Will candidates sit through a ten-minute AI interview to build a profile? Some will love skipping the awkward small talk. Others will balk at performing for a camera with no human on the other end. Fika's whole model rests on enough people choosing the first reaction over the second.
And employers have to trust it. A pool of pre-interviewed candidates only matters if hiring managers believe the assessment means something. That trust gets earned slowly, one good hire at a time, and lost fast the moment the agent waves through someone who interviews well and works badly. The first bad placement that goes viral on LinkedIn would cost more than any ad budget could buy back.
Four million dollars buys runway to find out. By the wider launch later this year, you'll know whether Fika built a new front door to the labor market or a very slick demo. Watch the retention numbers, not the press. If people keep their profiles live and employers keep coming back to the pool, the brothers were right about that near-miss hire. If not, the resume wins again, ugly and immortal.
