Abstract teal organic form suggesting interconnected learning paths and educational depth, warm forest green tones

We built Brainpathio because teachers deserve to know which misconceptions their students carry — before the unit test.

The hardest problem in adaptive learning isn't the algorithm. It's deciding what to measure. We decided to measure conceptual state, not completion rate.

The Founding Story

Jessica Tanaka spent her early career as a curriculum designer for a nonprofit educational publisher, working on STEM instructional materials for grades 3–8. She spent three years developing formative assessment frameworks — the kind that are supposed to help teachers identify misconceptions early, before errors calcify into deep conceptual misunderstandings.

The frameworks were good. The problem was implementation. Teachers were expected to review student work, identify error patterns, connect those patterns to underlying conceptual gaps, and adjust instruction in real time. For a department of three teachers, each running four sections, that meant analyzing the error patterns of over 400 students every two weeks. It didn't happen.

The moment that changed everything came during a classroom observation in 2023. Jessica watched a student answer 17 variations of the same algebra problem incorrectly — all wrong for exactly the same reason. The student confused negative exponent notation with "make the number negative." Every wrong answer was a signal. None of them were read. The teacher knew the student was struggling, but not why.

"I had spent three years helping curriculum teams design better assessments," Jessica says. "The assessments worked. The teachers understood what they were looking at. The problem was that there were too many students, too many problems, and the data lived in their heads or in a grade book. They needed a system that could read the signal they were already generating and surface it before the unit test."

She left the publisher in early 2024, completed a graduate certificate in cognitive science and learning analytics, and co-founded Brainpathio with Marcus Webb, who brought fifteen years of experience in learning science research. They built the first prototype in six months, ran an informal pilot with two Portland-area middle schools, and spent the next year refining the misconception taxonomy in collaboration with practicing STEM teachers.

Brainpathio received angel backing in March 2026. The team is currently four people, based in Portland, Oregon. We are small by design. We move slowly enough to get the pedagogy right.

What We Believe

Three principles that guide every decision we make.

Assessment should diagnose, not just rank.

Most assessment tools tell you how well a student performed relative to a rubric. We're more interested in what performance reveals about how a student understands a concept — and what they misunderstand about it. Ranking is the output; diagnosis is the point.

Teachers already know more than any algorithm — they need better tools to act on what they know.

Good teachers carry a working model of every student's conceptual state in their head. The problem is scale and time — one teacher, 120 students, 40 minutes per class. Brainpathio isn't replacing teacher judgment. It's making that judgment more legible, more systematized, and actionable before the test.

Adaptive learning is only as good as its conception model.

Every adaptive platform has a model of what students know. Most of those models are thin — right/wrong signals mapped to difficulty levels. We think the quality of the underlying concept model is the most important differentiator in adaptive learning, and it's the one thing most platforms treat as a secondary concern.

Talk to the Team

Questions? We're reachable.

Whether you're evaluating the platform or just want to talk about misconception mapping in STEM — we're happy to have that conversation.

[email protected] Contact Form