Semantic Search Engine with Sentence Transformers & Elasticsearch

Overview

Built a semantic search engine leveraging Sentence Transformers to generate text embeddings and Elasticsearch for scalable indexing and retrieval. Designed to handle both research papers (45+) and product datasets (5000+), the system supports natural language queries and retrieves the most relevant results using vector similarity search and ranking algorithms. Optimized indexing improved both accuracy and query performance, enhancing real-world search experiences.

Tech Stack

Python Sentence Transformers Elasticsearch Vector Encoding

Advantages