Introduction
Capstone project applies all the knowledge learned in the series to a real end-to-end problem.
Project requirements
Description
Project summary: Build e-commerce recommendation engine end-to-end. Two-tower retrieval + re-ranking + A/B testing + deployment.
Deliverables
| Item | Description | Weight |
|---|---|---|
| Code | Clean, documented GitHub repository | 30% |
| Report | Architecture decisions, results analysis | 30% |
| Demo | Interactive demo (web app or video) | 20% |
| Documentation | README, API docs, deployment guide | 20% |
Recommended pipeline
- Data Collection & Preprocessing: Collect and process data
- Model Development: Build and train the model
- Evaluation: Evaluation with appropriate metrics
- Optimization: Optimize performance and costs
- Deployment: Deploy to production
- Monitoring: Set up monitoring & alerting
Summary
Congratulations on completing the series! Apply your knowledge to real projects.