Kelly O. Jakpa

Where it started

Math is not hard. It is layered.

Kelly grew up in Nigeria. While other students memorized steps, he broke problems into their smallest pieces and rebuilt them until they made sense, first to himself and then to the people around him. By elementary school, classmates were already coming to him for help.

In middle school, a private tutor did something unusual: he taught Kelly ahead of his grade. While still in middle school, Kelly was learning high school mathematics, including calculus. Soon he was helping high school students himself.

Learning that far ahead showed him something he has built on ever since. Advanced mathematics is not a different kind of thinking. It is a combination of simpler skills, stacked in a particular order. Anything layered can be taken apart, and anything taken apart can be rebuilt.

Math is not hard. It is layered. And anything layered can be broken down.

In boarding school, among more than a hundred students, he became the one people came to when they could not understand. Not because he was the smartest person in the building, but because he could explain it. That is where he learned the difference between a student who has learned a rule and one who understands it. The first can apply it once. The second can apply it anywhere, and can explain it to someone else.

The road to Kmath

The road to Kmath

1. NigeriaBreaking problems into pieces, and helping classmates by elementary school.
2. Middle schoolTaught ahead of his grade, up to calculus. Helping high school students.
3. Boarding schoolThe student others came to when they could not understand.
4. United States, 2016Working with students and families in schools and communities.
5. Nine years running a learning centerThe method, refined one student at a time.
6. KmathThe method written down and built into software.

The method before the software

Find the step that broke

When a student got something wrong, Kelly did not start with more practice on the same problem. He worked backward through the steps underneath it until he found the one that had never become secure. Then he rebuilt from there.

1. Break down

   Take the skill apart into the smaller steps it depends on.

2. Find the missing prerequisite

   Check each step until the one that broke shows itself.

3. Rebuild

   Teach that step until it holds.

4. Reconnect

   Carry it back up to the skill the student was stuck on.

5. Check

   Come back later and see whether it stayed.

A wrong answer was a clue, not a verdict. The same mistake usually pointed to the same missing step. Once that step was found, the work that followed was short and specific instead of long and general.

United States, 2016

A pattern that kept repeating

Kelly came to the United States in 2016 and worked with students in schools and communities, with families from many languages, cultures and backgrounds.

Across all of them, the same pattern appeared. A child who looked broadly "behind in math" was rarely missing everything. Usually one earlier step, sometimes from years before, had never become secure, and everything built on it was unsteady. Students were not failing because they were incapable. Nobody had found the step that broke.

The parent problem

He saw a second problem too. The parents he met cared and wanted to help. But the report was in English, the meetings were in English, and the homework was in English. When information does not come in a language or a form a parent can act on, that parent is shut out of their own child's learning.

They were not uninvolved. They were excluded.

Years with students and families

What one student at a time taught him

Kelly spent about twenty years working directly with students, nine of them running a learning center. Year after year, the same lessons came back.

• A wrong answer is not the misconception.The answer shows that something broke. The pattern of mistakes shows what broke.
• Order matters.A skill cannot hold if the step under it is missing, so the repair has to start below the place where the student is struggling.
• Secure means it stays.A concept that works on Tuesday and is gone by Friday was not fixed. It has to be checked again later.
• More worksheets are not the answer.The right work, aimed at the missing step, does more than a bigger pile of the wrong work.
• Parents need a finding, not a score.A percentage tells a family very little. "This is the step your child is missing, and this is what happens next" is something they can act on.

His approach worked. Students improved, their confidence grew, and families trusted him. Children who had decided they were "bad at math" often lost that fear the moment the right explanation arrived.

The family dimension

The problem became personal

Kelly is also a parent. All four of his children are homeschooled, so the questions every parent asks became his own, every day, four times over: what exactly should this child learn next, and how do I know it has really stuck?

Four children also means four different starting points. One of his sons is ready to move faster than his grade, which showed Kelly the other side of the same problem. A child who is ahead needs the same precise map as a child who is catching up: which steps are secure, which come next, and in what order. Kmath is built for both, and for the child who is right on grade level too.

Building Kmath at home, inside his own family, kept one standard in front of him. It had to be something he would trust with his own children.

Why software?

One room at a time was not enough

The better Kelly got at the work, the more clearly he could see its ceiling. One student at a time. One session at a time. One location. The instinct that found the broken step could not travel beyond the room he was in.

And when he left the room, everything he knew about that child left with him. The next week, school had moved on, and the gap was still there. He was always one step behind the classroom.

The question he kept returning to was not how to reach more students. It was how to make the method itself repeatable, so it could run without him in the room.

1. One educator's repeatable method
2. Written down as diagnostic logic
3. Built into Kmath software
4. Available to many families

The goal was never to put Kelly on a screen and call it tutoring. It was to write the method down precisely enough that software could carry it: the diagnostic logic, the map of which skills depend on which, the layered breakdown of each skill, the weekly plan and the replanning, alignment with what the class is learning, the check-ins, the recalibration every four weeks, and the parent guide in the family's own language.

Kelly originated and developed the method. Kmath is the system that runs it.

From method to Kmath

Build the system, not a bigger tutoring schedule

Kmath runs the same loop Kelly once ran by hand, for every child, every week.

1. Observe

   The parent intake, then the child's own answers.

2. Diagnose

   21 questions show where understanding broke.

3. Map dependencies

   Each gap is traced to the earlier steps it rests on.

4. Build

   A printed weekly workbook aimed at those steps.

5. Learn

   Short practice on paper at home, most days of the week.

6. Check

   Short check-ins show whether each step held.

7. Recalibrate

   Every four weeks, a new assessment and a new plan.

Software lets that loop run again and again for each student, without Kelly personally teaching that child.

What experience became product decisions

Twenty years of lessons, built in

Each part of Kmath traces back to something Kelly saw work, or saw fail, with real students.

• Paper first

  Pencil on paper built understanding that screens did not. So the learning happens in a printed weekly workbook at the kitchen table, and the portal is kept for short check-ins.

• Exact steps, not broad labels

  "Behind in math" never helped a student. Finding the one missing step did. So Kmath names the specific skill, not a general grade level.

• Short and frequent

  Progress came from short, correct practice most days, not from one long session a week. So the plan is built around five to six short sessions a week.

• Explain it

  A student who could explain a concept to someone younger had truly learned it. So explaining to a sibling or a friend is part of the mastery check, not an extra.

• Real context

  "When will I use this?" deserves an answer. So each concept is tied to a real situation where it matters.

• A plan that moves

  No two students he worked with followed the same path. So the plan is rebuilt every four weeks from a new assessment.

From founder to company

Kmath has to work without Kelly in the room

The point of writing the method down was to remove the need for Kelly to deliver it himself. Kmath assesses the child, reports only the findings the evidence supports, builds the plan, produces the workbooks, checks progress, recalibrates, follows the classroom curriculum and keeps the family informed. All of that runs through the product and the way the company operates.

Building that company took two people. Tatiana T. James-Jakpa is Kmath's Co-Founder and Chief Operating Officer. She runs the operating side of the company around the system: family and customer operations, support, workflows, administration, reliability and day-to-day execution. As Founder and CEO, Kelly leads the product, the diagnostic and learning design, the technology and the company's direction.

Read Tatiana T. James-Jakpa's story

In his words

They were the reason

Kelly O. JakpaFounder & CEO
Creator and principal architect of Kmath