Why I built Math Ladder AI

The kitchen table

I am a professor of engineering. I have spent twenty-five years teaching university courses that lean hard on mathematics. And a few years ago, at my own kitchen table, in a highly competitive school district, I was struggling to keep math alive for my own two daughters.

Not for lack of trying. We did what every ambitious family around us was doing, the full enrichment circuit: Russian School of Mathematics, Mathnasium, Singapore Math workbooks at home. These are serious programs, run by people who care, and they help many families. But living inside that routine, week after week, here is what I watched happen in my own home.

The pace was set by the group, not by my child. When one of my daughters was ahead, she waited. When she was behind, the class moved on anyway. One room, one lesson, whatever each kid in that room actually needed that day.

The peer pressure kept building. For our kids, enrichment class was not an escape from ranking. It was one more place to be ranked, one more score to compare in the car ride home.

And slowly, the dislike crept in. A shrug in late elementary school. Resistance in middle school. Something close to dread by high school, where grade pressure stacks on top of enrichment homework on top of school homework. The subject I love most in the world was becoming the subject my children most wanted to escape. That is the quiet failure of the enrichment grind: a child can complete every worksheet and still lose the thing that mattered most.

The lecture hall

At the same time, I was watching the other end of the pipeline. As a professor, especially in the years since COVID, I have watched a growing share of incoming engineering students arrive unready for Calculus I. Bright, motivated, admitted students placing into precalculus, failing introductory math courses, fighting just to stay with their cohort. Colleagues across the country describe the same thing, and the national assessments (NAEP) tell the same story. It is not a local problem, and it is not the students' fault. Engineering and the sciences cannot be taught without math fundamentals, and I have had to watch too many students struggle through college because of gaps that quietly formed years earlier, back in middle school.

It took me embarrassingly long to see that I was watching the same problem from both ends.

What we already know works

The fix has been understood for decades: one-on-one tutoring built on mastery. A good tutor finds exactly what a student knows, teaches the next thing the student is ready for, and does not move on until it is solid. The research on this is old and settled. Students taught this way dramatically outperform students taught in a group. The problem was never whether tutoring works. The problem is the price. At a typical $60 to $100 an hour for a private tutor, or several hundred dollars a month at a center plus the weekly drive, real one-on-one teaching is out of reach for many families in this country, let alone the world.

What makes Math Ladder AI different

There are plenty of math apps. Most of them hand every child the same track for their grade and call it personalization. That is the classroom problem all over again, just on a screen.

The whole point of Math Ladder AI is that it has no fixed track. I used AI as my engineering partner to design and to code the expert system at the heart of this platform, an engine built around a single idea: meet the child exactly where they are, not where their grade label says they should be. One professor with AI at the keyboard can now build what used to take a funded company, and that is exactly how a family company can offer this at a family price.

That works in both directions. If your fifth grader is ready for seventh grade math, the ladder keeps climbing. Nothing caps them at grade level. They simply keep earning the next skill. And if your seventh grader is carrying gaps from fourth grade, the engine finds those exact gaps, teaches them without shame or drama, and coaches your child rung by rung back up to where they need to be. They are never labeled behind. They are just on their own rung of the ladder, moving up.

Honesty is a design principle here, so let me be precise: Math Ladder AI is not a person, and I will not pretend it is one. It is the part of great tutoring that software can genuinely do, done carefully. It maps what your child actually knows onto a graph of prerequisites, usually in one short sitting. It serves the next lesson only when your child is ready for it. It brings skills back for review right before they would be forgotten. It treats mistakes as the curriculum, re-teaching and retrying them until they are sealed. And it measures progress against your child's own yesterday, never against the kid at the next desk.

And when your child gets stuck on a doubt or carries a misconception, live AI steps in to tutor through it. The built-in coach explains the why behind any step in plain language, as many times as it takes, and Snap and Solve reads a photo of a hard homework problem and walks through the idea behind it. The expert system decides what your child learns next; the live AI helps them through the rough spots on the way. That is the division of labor, and it is an honest one.

Kindergarten through 12th grade, competition math through SAT, ACT, and AP. Every problem is generated fresh, from problem engines under ongoing internal validation - answers re-derived with a computer-algebra system and self-graded across hundreds of cases per topic. All of it for less per year than a typical single month of the programs we used to drive to.

Why it costs what it costs

I want to be straight about the price, because the price is part of the mission. When paid plans begin, Math Ladder AI will be $12.99 a month, $99.99 for a full year, or $149.99 a year for a family plan that covers all of your children. During our Founding Family beta, everything is free.

Here is what those dollars are for. This is not a venture with a big team, offices, franchises, and an advertising budget. It is a small, home-based, family-run company trying to deliver everything a large corporation would. The subscription pays the real bills behind the product: the servers and cloud services that keep the portal fast, the hosting and database fees, and the AI compute behind the live features. When Snap and Solve reads a photo of a homework page, or the built-in coach explains why a step works when your child is stuck, that answer runs on paid AI processing. Every one of those moments has a real cost, and as more families use them, that cost grows too.

The big enrichment chains charge several hundred dollars a month, and part of that price is buildings, staff, franchises, and marketing. We do not carry that overhead, so the savings go to you. If we do this right, your subscription covers the computing, keeps the independent problem audit going, and funds the next features. That is the whole business model.

The goal

The mission is bigger than my two daughters. It is to make excellent, mastery-based math teaching accessible to every family, across the country and across the world, and to raise the overall math IQ of a whole population without it costing parents and communities a fortune. What a good tutor does should not depend on what a family can pay.

I hope this effort stays true to its original goal, and helps as many people as possible.