GloZ’s Natural Language Resume Search System Built on Amazon OpenSearch Service — Part 1: Data Pipeline and Indexing | Amazon Web Services

GloZ rebuilt resume search for about 100,000 translators on Amazon OpenSearch Service.

The new pipeline uses Amazon Bedrock for cleansing and embedding generation, enabling hybrid keyword plus semantic retrieval.

Korean search quality was improved with the Nori analyzer, while HNSW vector indexing supports scalable performance.

The case shows how multilingual recruitment search can reduce ops overhead and improve candidate matching quality.