Threaded

Inspiration

A lot of educational software focuses on giving students more content. More videos, more practice questions, more notes, and more explanations.

The problem is that students do not always need more content. Sometimes they need to know what they are missing.

I have had situations where I could understand most of a topic but still get questions wrong because of one small concept I did not fully understand. Looking through an entire textbook or watching an hour of videos to find that one gap is not very efficient.

I wanted to build something that could figure out where a student's understanding breaks down and show them exactly what they should work on next.

What it does

Threaded is a personalized learning tool that builds a map of what a student understands within a subject.

A student can upload class notes, a syllabus, assignments, or practice questions. Threaded breaks the material into individual concepts and connects them based on their prerequisites.

The student can then take a short diagnostic assessment. Instead of simply giving them a score, Threaded identifies specific gaps in their understanding.

For example, a student studying calculus might know how to find a derivative but struggle with the chain rule. Threaded could recognize that the problem is not derivatives in general, but a specific prerequisite that is causing the mistakes.

It then creates a personalized learning path showing:

What the student already understands.
Which concepts they are struggling with.
Which prerequisite concepts they should review.
What they should practice next.
When they are ready to move on.

As the student completes questions, the map updates based on their performance.

How I built it

I built Threaded as a web application with an interactive concept map as the main interface.

The application takes uploaded educational material and uses AI to identify important concepts, relationships, and prerequisites. These concepts are represented as a graph rather than as a simple list of topics.

Students can then complete generated diagnostic questions or upload existing assignments and practice problems.

The system tracks performance at the concept level rather than only calculating an overall percentage. If a student repeatedly makes mistakes involving one concept, Threaded increases its confidence that the concept needs review.

The main workflow is:

Upload class materials or choose a subject.
Generate a map of the important concepts.
Complete a short diagnostic.
Identify gaps in understanding.
Follow a personalized learning path.
Complete targeted practice.
Update the learning map as understanding improves.

I also wanted the interface to make the student's progress visible. Instead of showing a generic progress bar, Threaded shows how individual concepts connect and which parts of the subject are becoming stronger.

Challenges I ran into

One of the biggest challenges was figuring out how to represent learning in a way that is actually useful.

A subject is not just a list of chapters. Concepts depend on each other, and a student can struggle with an advanced topic because they are missing something much earlier.

I also had to think carefully about how to determine whether a student actually understands something. Getting one question right does not necessarily mean that a concept is mastered, so Threaded needs to look at patterns across multiple questions.

Another challenge was making AI-generated questions useful rather than simply generating more generic practice problems. The questions need to target the specific concept that the student is struggling with.

I also wanted to avoid turning Threaded into another chatbot. The goal is not to have students ask AI for answers. The goal is to use AI to understand their learning process and help them decide what to do next.

Accomplishments that I'm proud of

I am proud that Threaded turns something that is usually difficult to see into something visual.

A student might know that they are getting questions wrong, but it can be difficult to understand why. Threaded tries to show the underlying structure of those mistakes.

I am also proud of the concept map. Instead of treating every topic independently, it shows how concepts connect to each other. This makes it possible to identify situations where a student is struggling with a topic because of a missing prerequisite.

Another part I like is that the system changes as the student works. The learning path is not fixed at the beginning. New performance data can change which concepts Threaded recommends next.

Most importantly, I wanted the project to focus on improving how students learn rather than simply giving them another place to consume educational content.

What I learned

I learned that personalized education is much more complicated than simply asking AI to generate a study plan.

A useful system needs to understand the structure of the material, estimate what a student knows, and connect mistakes to the concepts that caused them.

I also learned that visualizing information can sometimes be more useful than adding more information. A student may already have access to everything they need to study. The difficult part can be figuring out where to start.

Building Threaded also made me think more about how AI can work alongside traditional education. Instead of replacing teachers or doing the work for students, AI can help students and teachers understand where learning is happening and where it is breaking down.

What's next for Threaded

I would like to expand Threaded into a system that can work across an entire semester.

Students could connect their assignments, quizzes, and exams throughout a course, allowing Threaded to maintain a long-term model of their understanding.

I would also like to build a teacher dashboard. Teachers could see which concepts an entire class is struggling with without needing to manually analyze every assignment.

Another feature I would like to add is collaborative learning. If several students are struggling with the same concept, Threaded could organize study groups or recommend students who could help each other.

Eventually, I want Threaded to work across different subjects and education levels, from high school math and science to university courses.

My goal is to make education feel less like moving through the same textbook at the same pace and more like having a clear path through the material based on what each student actually understands.