PathwayAI

Smart Course Selector AI Agent

Contributors: Rishikesh Kakde, Gayatri Gattani, Rutuja Jangle, Muskan Dhingra

Welcome to PathwayAI, an intelligent AI agent that simplifies the course selection journey for students through a clean, modern UI powered by OpenAI's Agent SDK.


Project Overview

Objective

An intelligent, student-centric platform that transforms the often overwhelming course selection process into a smooth, personalized experience. This system empowers students to:

  • Explore degree programs and professional pathways aligned with their long-term career goals (e.g., Data Scientist, Product Manager, UX Designer).
  • Receive goal-driven course recommendations tailored to their academic background and credit load preferences (full-time/part-time).
  • Gain clarity on prerequisites by conversing with the agent to understand which foundational courses are required for a desired major or advanced course.
  • Navigate real-world limitations like limited seat availability and high-demand course slots with intelligent alternatives and dynamic planning.
  • Engage in meaningful conversation with a chatbot assistant to resolve doubts, clarify academic rules, and make informed choices in real-time.

Why This Matters

Course selection is often stressful due to:

  • Uncertainty in aligning choices with career paths.
  • Limited knowledge of prerequisites or credit rules.
  • Fast-filling classes and limited availability.

    System Architecture & Methodology

  • Dual-Agent Design:

    • Chatbot Assistant (GPT-4o-mini): Engages students in free-form conversation to answer questions about course selection, credits, prerequisites, etc.
    • Recommendation Assistant (GPT-4): Accepts structured student inputs (e.g., desired profession, major) and responds with tailored course recommendations considering prerequisites, availability, and credit load. The recommendations are shown in the format of a Visual roadmap of course progression vs career goal.
  • OpenAI Assistant SDK:

    • Integrated using Python scripts (Agents_Chat.py) and Jupyter Notebooks (CourseSelectorAgent.ipynb, UIListCoursesAgent.ipynb).
    • Threads and runs are managed via the OpenAI Assistants API, supporting persistent, contextual conversations.
  • Model Choice:

    • GPT-4o-mini for lightweight, fast responses in chatbot interactions.
    • GPT-4 for high-quality, goal-aligned, and reasoning-heavy course recommendations.

Frontend (UI)

  • Built with: TypeScript, using modern component-based architecture.
  • Hosted on: Vercel for fast deployment and seamless CI/CD.
  • Features:

    • Clean layout to host and switch between the two AI agents.
    • Input forms for structured career/course queries.
    • Live chat interface for interacting with the assistant.
  • Key Pages:

    • Landing Page
    • Modern, animated hero section
    • Featured degree programs
    • How it works section
    • Quick access to course exploration
    • Explore Page
    • Course catalog with filtering
    • Real-time seat availability
    • Course details and prerequisites
    • Interactive selection system
    • Degree Planning
    • Visual roadmap of courses
    • Semester-by-semester planning
    • Progress tracking
    • Prerequisite visualization
    • Interactive Elements:
    • AI Chatbot
    • Course recommendations
    • Career guidance
    • Prerequisite checking
    • Real-time assistance
    • Course Selection
    • Seat availability indicators
    • Prerequisites validation
    • Interactive cards with detailed info
    • Progress tracking
    • Smart Features:
    • Availability Tracking
    • Real-time seat numbers
    • Filling rate visualization
    • Automatic status updates (Open/Limited/Full)
    • Career Alignment
    • AI-powered course suggestions
    • Career path optimization

    - Skill gap analysis

Knowledge Base

  • Located in the Knowledge Base/ directory.
  • Contains university course PDFs for all programs offered by Luddy School to ground the agents with accurate prerequisite and curriculum knowledge.
  • Serves as embedded context for better reasoning and recommendations.

Future Enhancements

  • Course review aggregator - Integrate anonymized course reviews and ratings to help students make informed choices.
  • Timetable conflict detection - Detect overlapping course times and recommend conflict-free alternatives automatically.
  • Integration with university enrollment APIs - Integrate with institutional APIs to show real-time course availability, registration deadlines, and syllabus links.
  • Learning Style-based Recommendations - Recommend courses based on preferred learning methods (e.g., project-based, lecture-heavy, hands-on labs).

Built With

+ 28 more
Share this project:

Updates