MathGPT Careers
No funded vacancies today, the five roles we would fill first, and what makes a work sample stand out.
The short version
- There are no funded vacancies posted right now, and we would rather say so than run a permanent fake job board.
- We do read every message, and we keep the good ones. Five roles are listed below with what the work actually is.
- Show evidence, not enthusiasm: what you built, what you decided, and how you knew it worked.
- Write to [email protected] with one work sample and one paragraph about it.
The honest position
This is a small team, and there is no open vacancy with a budget attached to it today. Plenty of career pages hide that behind a permanently open "general application" form; we would rather put it in the first paragraph. What is true is that we read everything sent to [email protected], we keep work samples that impress us, and the roles below are the ones we would fill first if we could fill one. If that is enough for you to write, we would genuinely like to hear from you. If you need a live posting, come back later, because this page changes when the position does.
What the work is really about
The difficult part of this product is not producing an answer. Any model can do that. It is producing something that leaves a student able to do the next problem alone.
- Deciding what a learner should see first when they are stuck and slightly panicked.
- Making a wrong AI answer easy to catch instead of easy to trust.
- Writing mathematics that is exact and still readable by a fourteen-year-old.
- Building tools that run in a browser, fast, with no account and no tracking.
- Keeping every page usable by someone who cannot use a mouse.
The roles we would fill first
Each of these is real work that exists on the site today, with pages you can go and inspect before you write. The status column is the same for all of them for now, and we are not going to dress that up. Read the areas, pick the one where your evidence is strongest, and write about that one rather than all five.
| Role | What you would actually do | Status today |
|---|---|---|
| Mathematics editor | Write and check guides, worked examples, and practice sets across /learn and /practice, and find the ones that are wrong. | Expressions of interest. No funded vacancy. |
| Learning designer | Turn a syllabus into a path a student can follow, across courses, AP guides, and the study planner. | Expressions of interest. No funded vacancy. |
| Front-end engineer | Build the browser-side tools: graphing, geometry, matrices, 3D surfaces, and the notebook. They run without a server call, and that is the constraint. | Expressions of interest. No funded vacancy. |
| Accessibility specialist | Test with real assistive technology, fix what fails, and own the honest status table on the accessibility page. | Expressions of interest, and the gap we feel most. |
| AI evaluation | Design tests that catch a wrong step in the middle of a correct-looking solution, not just a wrong final answer. | Expressions of interest. No funded vacancy. |
What a strong work sample looks like
We care much less about where you worked than about what you decided and how you knew it was right. The best samples we have received were small and specific: one lesson, one component, one evaluation, explained honestly including the part that did not work. The worst were long lists of technologies. If you write one paragraph and attach one thing, make the paragraph about the decision you are proudest of and the evidence behind it.
Answer these five questions about your sample
Any format is fine: a document, a repository, a lesson, a video. The questions matter more than the format.
- What was the problem for the person using it?
- What constraints were you working inside: time, budget, curriculum, browser?
- Which part did you personally do?
- What was the hardest trade-off, and which way did you go?
- How did you find out whether it worked, and what did you change afterwards?
Things that stand out here specifically
These are the signals that make us read a message twice.
- You found a real error on this site and explained it precisely.
- You can show a piece of teaching that made a hard idea land.
- You have built something that works well without JavaScript frameworks or a tracking script.
- You have tested something with a screen reader and know what actually broke.
- You can say clearly when you were wrong about something and what changed your mind.
How to write, and what happens next
Send a short note to [email protected]. Put the area you are interested in in the subject line, write one paragraph about one piece of work, and attach or link to it. You do not need a formal covering letter and you do not need to flatter the product; if anything, telling us what is weak about it is a better opening. Because there is no open vacancy, do not expect a fast reply, and do not read silence as a judgement of your work. We keep what we like and we come back to it when there is a role. In the meantime, the about page explains how the site is actually built, and it is the fastest way to work out whether you would enjoy this.
Show evidence, not enthusiasm
The strongest work samples we receive are small and specific: one lesson, one component, one evaluation, explained honestly including the part that did not work. Say what the problem was for the person using it, what constraints you had, which part you personally did, the hardest trade-off you made, and how you found out whether it worked. For mathematical and teaching work, show how you checked accuracy and whether the learning transferred. For engineering and design, show what you did about accessibility and failure cases.
Questions people ask about this page
Are there any open positions right now?
No funded vacancies are posted today. That is the plain answer, and this page is updated when it changes. Expressions of interest are still read and kept, and the five areas above are the ones we would hire into first.
Do I need a mathematics degree?
Not for every role. Editorial and learning design roles need real depth in the mathematics you would be writing about, however you got it. Engineering and accessibility roles need depth in their own craft plus enough comfort with math to know when something looks wrong. Teaching experience counts heavily for all of them.
Is this remote?
Yes. The team works remotely and asynchronously. That means writing clearly matters more here than it does in an office, which is one reason we ask for a written work sample instead of a long interview loop.
What is the fastest way to get our attention?
Find something wrong on this site and explain it properly: the page, the exact line, why it is wrong, and what it should say. That single email tells us more about how you think than any CV, and it is useful to us whether or not it leads anywhere.
Contact MathGPT.now
For questions related to this page, email [email protected]. Include the page address and enough non-sensitive detail to reproduce the question or problem.