🤝 Top Collaborators
🤓 Latest Submissions
Personalized AI Learning Path Recommender
Finding the right learning path online can be difficult because platforms such as Coursera, Udemy, and edX contain thousands of courses, making it challenging for learners to identify what to study next and how to progress from beginner to advanced levels. Our Personalized AI Learning Path Recommender addresses this problem by combining Natural Language Processing, machine learning, semantic similarity, and conversational AI into a single interactive platform. A learner can describe a topic, skill, or learning goal in natural language, and the system analyzes the input to identify relevant courses from a large collection of course reviews. The recommendation engine preprocesses review text using a custom text-cleaning pipeline and converts it into TF-IDF representations using up to 100,000 features and 1–3 word n-grams. A NearestNeighbors model using cosine distance retrieves similar course reviews, while a Logistic Regression classifier predicts the relevant course category. These results are combined through course-aware candidate re-ranking to produce the Top 10 personalized recommendations. Beyond course recommendations, the platform provides conversational AI guidance through supported LLM providers such as Gemini, Groq, and OpenRouter. The AI can use the recommendations to generate structured three-phase learning roadmaps containing milestones, estimated completion times, focus areas, and topic tags. The application also includes an interactive React dashboard with chat history, progress checklists, recommendation cards, roadmap visualization, authentication, provider/model selection, and API-key configuration. The backend is implemented using FastAPI and Python, while the frontend uses React, TypeScript, Vite, and Tailwind CSS. The system can operate with external AI APIs or use its built-in fallback mode when an API key is unavailable. The project also supports batch prediction and evaluates the recommendation pipeline using Recall@10.
27 Sep 2026
