1. ingestion pipeline (scheduled 1/d)

• fetch raw stock data
• generate features
• push features in feature store

2. model training pipeline

• pull features/target from feature store
• train model
• push model to model registry

3. predict pipeline

• pull historical data
• generate features for next day
• pull model
• predict next day stock price from model

4. streamlit app (WIP)

• show features
• show prediction
• show difference between target & predictions