What we learned building a job platform nobody asked for

"Another job board" is a fair reaction, and it was ours too. The category is saturated, the incumbents are enormous, and the failure mode is well documented. We built it because a specific complaint kept surviving every attempt to talk ourselves out of it.

The complaint that would not go away

Engineers in Europe filter for remote work and get a list that is mostly unavailable to them. They cannot tell which listings are real without opening each one. They apply into silence and learn nothing from it. Every part of that is a data problem, and none of it is a problem the incumbents are motivated to solve — a board paid by volume has no reason to show you fewer, better roles.

What turned out to be true

• The filter was the product. We thought the score was. Everything downstream depends on the corpus being genuinely EU-eligible, and that took far longer than the scoring did.
• Going to the source is unglamorous and decisive. Reading company career pages rather than reposts means the eligibility text is intact when we parse it. It is also the only part of the system that cannot be shortcut.
• Ambiguity is the whole job. Two-letter state codes, cities that exist in three countries, "Remote" meaning four incompatible things. Every one of these is a special case, and the special cases are the moat.
• Saying less is more persuasive. Our early copy claimed freshness we had not verified. Cutting the claim to what we could defend — sourced direct from career pages — made the pitch shorter and considerably harder to argue with.

What we got wrong

• We treated worldwide-eligible roles as a lesser category early on, which quietly cost us real listings that EU engineers could have taken.
• We wrote strategy documents describing how a filter behaved without re-reading the filter. One of those descriptions was wrong for months and shaped decisions before anyone checked it against the code.
• We built scoring before we had any way to validate it. That is still true, and we now say so on the score itself rather than hoping nobody asks.

Where it stands

Roles sourced direct from company career pages across the EU tech market, filtered for genuine EU eligibility, scored against your profile for interview likelihood. The score has no outcome validation yet and we say so. The filter still gets edge cases wrong and we would like to hear about them.

If you are applying into silence from somewhere in Europe, this was built for you specifically. That is the whole pitch.