Import data, search and organize

1. Setup cloud bucket integration
2. Create a dataset
3. Data curation
4. Zero-shot labeling
5. Import ground truth annotations
6. Import model predictions as pre-labels
7. Import model predictions for model error analysis

Label data

1. Setup ontology (task design)
2. Setup labeling project
3. Use internal team or your own vendor
4. Use Boost: Data labeling service

Train models and diagnose errors

1. Overview
2. Model error analysis
3. Train models