Docere: A Persistent-Memory System for Personalized Coding Education

CS Education - MN Case Study

Coding education is failing not because students can’t learn, but because our systems can’t remember.

In places like Minnesota, access to computer science is already limited. But even when students do get exposure, most modern learning platforms—AI tutors included—treat every session as if it’s the first. Misconceptions are forgotten, progress doesn’t compound, and instruction resets instead of building forward. Learning becomes repetitive, shallow, and discouraging.

Education is cumulative. Coding especially so. Early misunderstandings don’t disappear—they compound into larger failures. Human tutors naturally adapt because they remember. AI systems largely do not.

Docere exists to fix this.

Tree Metaphor

Learning in Docere is modeled as a tree:

• Shared fundamentals form the trunk

• Individual decisions create branches

• Mastery accumulates as growth over time

No two learners follow the same path—and Docere doesn’t force them to.

By persisting structured memory across sessions, Docere enables learning that compounds, adapts, and respects the individuality of how people think.

Education is not a sequence of prompts.

It is a process shaped by memory.

Docere is built to remember.

Current

At the time of writing this Docere has been deployed in early production environments and currently supports over 1,000 learners through a vendor partner. Initial results indicate improved retention and learner confidence compared to stateless tutoring systems.

The system has raised over $10,000 in early funding, validating both technical feasibility and market interest.