Using a real-world Walmart sales dataset, the app forecasts future sales of a selected product for a number of days (user-defined) at different store locations across the US — starting from the last date on which the model was trained.

Train & Predict Pipelines — handles end-to-end ingestion, pre-processing, training, and model selection with just a single trigger.

Recursive Forecasting — the model generates multi-day predictions by feeding previous predicted values as inputs for future ones.

Clean API design - simple easy-to-read FastAPI code.

User-Friendly Interface - clean and straightforward React UI.