Advanced · 3–4 weeks
Semantic Search over Product Catalog
Natural language product search: 'comfortable running shoes under $100'. Hybrid search combining keyword filters with embedding similarity. Re-rank top results.
Why this project
Bridges traditional e-commerce and modern AI search — great for senior+ interviews.
Tech stack
Java 17Spring Boot 3Spring AIElasticsearch + dense_vector or pgvectorPostgreSQL
Skills demonstrated
- Hybrid search
- Embedding pipelines
- Re-ranking
- E-commerce integration
Interview talking points
- Hybrid vs pure vector search
- Embedding model choice
- Cold start for new products
- Measuring search relevance
Problem statement
Natural language product search: 'comfortable running shoes under $100'. Hybrid search combining keyword filters with embedding similarity. Re-rank top results.