Wivme
Apply

Wivme · Founding team

Technical Co‑Founder

Every week another startup ships a personalized AI tutor. We are not building one. Answering the student is the easy half of learning, and on its own it is close to useless, because none of these tools can tell you whether anything actually stuck.

Wivme does the hard half. It reads what a student genuinely wrote, finds the exact idea they have wrong, fixes it, and proves the fix held. That is a real ML and systems problem, not a wrapper on someone else's model, and it is the part that turns into a company. We are looking for the person to own it.

ApplyRead the role first
01

What we're building

The most powerful result in education is forty years old. In 1984 Benjamin Bloom showed that a student with a one-on-one tutor outperforms a normal classroom by two standard deviations, better than 98 percent of their peers. It has been useless the whole time, because you cannot give forty children a tutor each. AI finally makes it possible, and almost everyone is racing after the same easy half: a chatbot that answers questions.

We are building the hard half. A student produces a real answer, in their own words or their own working, never A, B, C, or D. We read the reasoning itself, catch the precise misconception, bring it back right before it is forgotten, and re-check to prove the gap closed.

Spotting that a student is wrong is twenty years old. Proving which fix actually closes which misconception, and getting sharper at it with every student, is unsolved. Nobody owns that, and it is the most valuable thing there is to build in education. That is the frontier we are walking into.

02

The role

You own the technology end to end and lead the engineering. This is the earliest and highest-ownership technical seat at Wivme: you make the architectural calls, set the bar, and decide how the whole thing gets built.

You will be judged on one thing, whether it works: in a real classroom, on a real cheap phone, for real students. Not on how clean the diagram looks. And you will hire and lead the engineers who build it with you.

03

What you'll own

  • The core system

    How a learner's activity is captured, trusted, and turned into what every screen shows. The system of record the whole product runs on.

  • The intelligence layer

    Turning a page of a student's real working into a precise, trustworthy read of what they understand, and holding the exact line where a model may decide and where it must never.

  • The student experience

    A fast, forgiving app on cheap phones and weak networks, where a student's work is never lost.

  • Reliability and trust

    Judgments accurate and honest enough that a teacher acts on them and keeps trusting them. The difference between a tool teachers rely on and one they quietly ignore.

  • Scale and cost

    Running this for pennies per student at the scale of a country, without the model bill becoming the business, and with students' data handled responsibly.

  • The team

    The engineering culture, the hiring bar, and the first engineers you bring on and lead.

04

What we look for

The work is the filter, not seniority or pedigree. But this is not something you build in a weekend, so:

  • You have architected systems that ran in production, with real users, and lived with what broke at 2am.
  • You go deep across the stack, from the data model to the last mile on a slow phone, and you can decide where a model belongs and where it must never be.
  • You use AI to move fast and still own every judgment. The taste and the calls are yours; the model is just faster hands.
  • You can lead a team: set the bar, make the hard calls, and bring people up to build with you.
  • You are drawn to hard, unsolved problems, and bored by CRUD apps and thin wrappers on someone else's model.
05

How to apply

The form is short. It asks for your best real work, and a swing at a few of the actual problems above.

Start your application

The founder reads every one personally.