The rapid growth of digital content platforms demands efficient and personalised tools for item recommendation. This paper presents Podsage AI — a Flask-based web application that enables end-to-end automation of the recommendation workflow. The system accepts user-item interaction datasets uploaded in CSV format, automatically performs preprocessing including encoding and normalisation, and simultaneously trains four widely-used recommendation algorithms: User-Based Collaborative Filtering, Item-Based Collaborative Filtering, Matrix Factorisation (SVD), and a Neural Collaborative Filtering (NCF) model. Precision@K and RMSE scores are computed for each model and the best-performing algorithm is identified and reported. Results are visualised through a comparative bar chart rendered within the web interface. The system is evaluated on two benchmark datasets — the MovieLens 100K dataset and the Amazon Product Reviews dataset. The proposed platform lowers the barrier for non-expert users to deploy recommendation engines and provides a reproducible, extensible baseline for automated model selection research.
Swathi G (Sat,) studied this question.