The company brain that maintains itself
Mention turns the documentation, meetings, and chats a team already produces into a living, self-maintaining knowledge base — so the right understanding reaches the right person before it gets expensive to be missing.
Why I built it
Mention started as a private glossary. At every engineering job, I kept a running list of the terms people threw around in standups and design docs, each paired with what it actually meant. It was a small habit that quietly did most of the heavy lifting of onboarding: once you know what the words mean, the rest of a system mostly explains itself.
The usual story about institutional knowledge is that it walks out the door when people leave. That's real, but it undersells something more common and more costly: the people still in the building are quietly operating on slightly different definitions of the same words, and nobody notices until a decision built on the mismatch turns out to be wrong. A surprising number of "communication problems" are really definition problems.
Documentation is supposed to solve this, but it rots the moment you ship it. Updating it is the kind of task you can deprioritize forever, because nothing breaks the day you skip it. Mention exists to close that gap: a glossary, playbook, and learning experience that rebuild themselves from your real source material instead of asking anyone to maintain a document by hand.
I've been a machine-learning engineer since 2017, and I've spent my whole career inside startups — which means that alongside the ML work, I've usually been the one holding the rest of the system together too. I've shipped review-mining systems that distilled millions of customer reviews into something a team could actually act on, identity-resolution and ad-targeting systems, and NLP automation for cold sales. The thread running through all of it: turning messy human language into structure a computer — and a person — can work with.
I'm largely self-taught, without a traditional computer-science degree, and I think that's a big part of why this problem grabbed me. Teaching yourself everything makes you pay close attention to how learning actually works. Mention is the system I always wished I had: the private glossary I kept rebuilding at every job, turned into something a whole team can share.