Pluggable Semantic and Lexical search for Large Datasets
Praminda Jayawardana
Like this project
Posted Jul 21, 2025
Deliver personalized, context-rich search results in under 200ms with In built pipelines for indexing external dataset from multiple sources.
Likes
0
Views
2
Timeline
Jan 1, 2024 - Mar 30, 2024
AI-Powered Semantic Search for Intuitive Product Discovery and Increased Sales.
A pluggable search service which leverages Semantic Search and Generative AI to provide contextually relevant, semantically rich search results. Increase engagement, conversions, and revenue with fast & intelligent search.
Requirements
Fast and easy data indexing of different external (customer) datasources (PG, S3 buckets, Web sites etc) into GDN vector store.
Indexing pipeline should be able to process both structured (json, html etc) and unstructured (text, pdf etc) text.
Quick search results based on semantic meaning of user query.
Support for advance search filters.
Asynchronous AI answer generation for the user query based on the search results
Contribution
Leading the project from finalizing the requirements, designing & architecting to the deployment
Implementation of REST APIs and business logic for the search.
Implementation of AI answer generation.
Making indexing process highly customizable to cater different types of data (Wikis, Articles, product catalogues etc)
Improving and tuning performance based on different tests.
Documenting, deploying (with CICD) and maintaining the product in different kubernetes environments.