Developing a Scalable AI Search System with Seamless IntegrationDeveloping a Scalable AI Search System with Seamless Integration
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Hikari's avatar
Hikari Builds
• Jun 4
Enterprise AI Search System with Embeddings, Pipelines & Widget Builder I worked as a Full Stack Engineer at Gloo, contributing to the architecture and development of an AI-powered discovery widget designed for large-scale content platforms.
The product is an embeddable search system that enables semantic content discovery across podcasts, sermons, articles, and other media, allowing organizations to surface relevant content through intent-based search rather than traditional keyword matching.
I helped design and build the system end-to-end, working across frontend, backend, and data infrastructure. On the frontend, I developed a modular widget and configuration system using Next.js, TypeScript, and modern UI tooling, enabling partners to integrate the search experience with a single script tag.
A key part of the project was the widget builder interface. I built a multi-step configuration system with live preview, allowing non-technical users to customize layout, data sources, API keys, and embedding behavior in a simple and intuitive workflow.
On the backend and data layer, I implemented ingestion pipelines to process large volumes of publisher content via RSS feeds. This included generating embeddings, enriching metadata, and preparing data for semantic retrieval, enabling domain-specific AI search across diverse content types.
The search system itself was built using a hybrid approach. I combined vector-based semantic search using Weaviate with high-speed autocomplete and indexing via Typesense, resulting in accurate intent-based results alongside instant query suggestions.
I also contributed to system reliability and observability by integrating analytics and monitoring tools, enabling detailed tracking, debugging, and performance optimization across deployments.
The result is a scalable, production-grade AI search platform that can be embedded into external sites with minimal effort, providing powerful discovery capabilities backed by modern AI infrastructure.
This project highlights my ability to design and build complex AI systems that combine frontend experience, backend architecture, and data pipelines into a cohesive, high-performance product.
Completed work
Backend DevelopmentFrontend DevelopmentFull Stack DevelopmentGraphQLLangChainNext.js
Post image
Post image
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started