NeuroBridge

NeuroBridge is an explainable AI adaptive learning platform that uses personalized recommendations, progress tracking, and faculty oversight to help neurodivergent undergrads manage their coursework.

Comment

Inspiration

Many neurodivergent university students find difficulty in learning not because they are not capable of it, but because in traditional learning environments, there are no considerations for their special needs. It is, in most cases, not designed to accommodate them. It can start to feel overwhelming for the students to manage their coursework, understand what their professors expect of them and maintain steady learning progress all at the same time.

We wanted to create a platform that bridges the gap between students and professors by providing adaptive support platform while keeping faculty involved in the learning process.

It was important to us that we create something that's not meant to replace educators but instead create a system that empowers both students and teachers through explainable AI, personalized recommendations, and actionable learning insights.

This vision led to the development of NeuroBridge.

What it does

NeuroBridge is an AI powered explainable and adaptive platform designed especially for neurodivergent university students.

Students Perspective

  • Help finish daily learning check-ins.
  • Get personalized recommendations.
  • Monitor assignment progress status.
  • Use analytics to track learning patterns and trends.
  • Access alternative learning paths when needed.

Educators Perspective

  • Create and Manage assignments.
  • Set timelines and priorities.
  • Provide assistance specific to the current assignment.
  • Monitor student behavior, engagement and progress.
  • Get alerts about students who may need additional help or guidance.

AI perspective

  • Natural Language Processing (NLP)
  • Anomaly Detection
  • Progress Analytics
  • Adaptive recommendation Engine
  • Explainable decision support

How we built it

How we brought our idea NeuroBridge to life.

1. Building the Adaptive Learning Engine

We began by designing a structured learning model centered around three key inputs from user:

  • Specified priorities and assignment information from teacher.
  • Student daily check-ins
  • Past trends in student learning progress Instead of relying completely on an LLM, we built an explainable learning engine that uses recommendation algorithms, rule-based reasoning and anomaly detection to generate personalized learning support for students.

This engine evaluates:

  • Assignment deadlines
  • Teacher specified priorities
  • Difficulties and challenges reported by students
  • Previous assignment completion patterns
  • Past trends in student confidence level and engagement metrics

Using these signals, the system generates a Next Best Action recommendation for each student while keeping educators in control of final decisions.

2. Using NLP to Understand Educational Context

To better understand inputs from student and teachers, we incorporated Natural Language Processing (NLP).

The NLP pipeline analyzes:

  • Student daily check-ins
  • Student comments explaining the challenges they are facing
  • Teacher assignment notes
  • Teacher recommendations to guide students

The system identifies:

  • Relevant courses
  • Assignment references
  • Learning difficulties
  • Support needs
  • Urgency indications

For example, when a student mentions:

"I have my Data Structures lab tomorrow and I'm struggling with graphs."

the NLP engine extracts:

  • Course: Data Structures
  • Deadline urgency: Tomorrow
  • Difficulty area: Graphs

These signals are then passed into the recommendation engine.

3. Developing the Recommendation System

The core of NeuroBridge is the recommendation engine that suggests what the student should work on next.

Each assignment receives a dynamic score that's based on the following criteria:

  • Teacher-specified priority
  • Assignment deadline
  • Student check in signals
  • NLP detected course references
  • Historical learning patterns of students

The highest scoring assignment becomes the student's next recommended activity. Most importantly, we made sure the recommendations remain transparent and clearly explainable. Why a particular recommendation was generated is crucial information for the student as well as the teacher.

4. Building the Anomaly Detection System

To help professors identify students who may need additional support, we developed an anomaly detection section.

The system will continuously monitor:

  • Repeated skipped assignments
  • Declining confidence levels
  • Low completion rates
  • Excessive task completion times

When unusual patterns are detected, NeuroBridge generates alerts to teacher and assigns an attention level:

  • Routine Support
  • Monitor
  • Intervention Recommended

This allows teachers to intervene early before academic difficulties escalate.

5. Designing the Teacher Dashboard

We created a teacher-focused dashboard that provides a concise overview of student progress.

The dashboard includes:

  • Active assignments
  • Assignment completion rates
  • Learning trends
  • Time spent on coursework
  • Anomaly alerts
  • Recommended support strategies

Teachers can also create assignments, set priorities, add support notes, and monitor progress across multiple students.

6. Creating the Student Experience

The student experience was intentionally designed to be simple, supportive, and non-judgmental.

Students:

  • Complete a daily check-in
  • Receive personalized recommendations
  • Track assignment progress
  • Record completion times
  • View learning analytics

Instead of highlighting weaknesses, NeuroBridge focuses on actionable next steps and positive learning support.

7. Visualizing Learning Analytics

To help both students and educators understand learning patterns, we integrated interactive analytics and dashboards.

The platform visualizes:

  • Assignment completion trends
  • Time spent per course
  • Productivity patterns
  • Confidence trends
  • Completed vs skipped work

These visualizations transform raw learning data into actionable educational insights.

8. Building the Full Stack Application

Frontend tools used:

  • React
  • TypeScript
  • Tailwind CSS
  • Recharts
  • Vite

Backend tools used:

  • Python to create the AI algorithms and Anomaly detection
  • Supabase- database to store

The frontend and backend work together to provide real-time recommendations, progress monitoring, and teacher support tools.

9. Human-Centered AI Design

A major design principle throughout development was ensuring that AI supports educators rather than replacing them.

NeuroBridge follows a human-in-the-loop approach where:

  • Teachers remain the final decision-makers.
  • AI recommendations are explainable.
  • Student data is used responsibly.
  • Support strategies are transparent.
  • Educational decisions remain accountable.

This creates a system that is both practical and trustworthy for real educational environments.

Architecture Overview

Teacher Inputs
        +
Teacher Notes
        +
Student Daily Check-In
        +
Assignment History
        ↓
Natural Language Processing
        ↓
Adaptive Scoring Engine
        ↓
Recommendation System
        ↓
Anomaly Detection
        ↓
Student Recommendation
        +
Teacher Dashboard
        +
Learning Analytics

Through this architecture, NeuroBridge delivers personalized learning support while maintaining educator oversight, transparency, and accessibility for diverse learners.

Challenges we ran into

Some of the biggest challenges included:

  • Balancing personalization with explainability.
  • Designing recommendations that remain supportive rather than discouraging.
  • Synchronizing assignment data, recommendations, analytics, and progress tracking.
  • Ensuring faculty remain involved in educational decisions rather than being replaced by automation.
  • Building a recommendation system that adapts to changing student needs.

Accomplishments that we're proud of

  • Built an explainable AI recommendation engine rather than relying solely on a chatbot.
  • Created a human-in-the-loop system that keeps faculty in control.
  • Implemented anomaly detection to proactively identify students who may require support.
  • Developed a complete workflow connecting assignments, recommendations, analytics, and progress tracking.
  • Designed a platform specifically focused on supporting neurodivergent university students.

What we learned

Through this project we gained experience in:

  • Natural Language Processing
  • Recommendation System Design
  • Anomaly Detection Techniques
  • Frontend and Backend Integration
  • Educational Technology Design
  • Explainable AI Principles

We also learned that effective educational AI is not just about generating recommendations, it is also about generating recommendations that are understandable, actionable, and trustworthy.

What's next for NeuroBridge

Our future roadmap includes:

  • More advanced NLP for deeper understanding of student and faculty inputs.
  • Predictive learning analytics to identify academic risks earlier.
  • Personalized study planning and scheduling.
  • Integration with existing university Learning Management Systems (LMS).
  • Expanded accessibility features for different neurodivergent learning profiles.
  • Real-time faculty intervention recommendations.
  • Longitudinal learning analytics across semesters.

Ultimately, we hope NeuroBridge evolves into a comprehensive adaptive learning platform that helps neurodivergent students thrive academically while strengthening collaboration between students and educators.

Built With

Share this project:

Updates