top of page

Swiftype to Elasticsearch Serverless Migration For An Airline Group

Google Cloud Data Analytics.png
Google Cloud Premier Partner.png
Elastic Partner Reseller.png

Client

The Client is a major American airline group modernizing its enterprise search infrastructure and multi-domain content strategy.

Project Context

The project successfully migrated a legacy Swiftype instance to a modern Elasticsearch Serverless architecture to eliminate technical constraints and reduce operational costs. The architecture utilized a hybrid ingestion layer—integrating the Elastic Open Web Crawler with custom Python scripts for Contentstack- to achieve a unified, high-precision search experience across 5+ web domains and five global languages.

Project Objectives

- Replace managed legacy systems with a serverless architecture for better scalability.
- Leverage advanced ranking algorithms to achieve higher precision and lower search latency.
- Support a hybrid ingestion model for both legacy web pages and new Headless CMS content.

Challenges

- Overcoming the limitations and high operational costs associated with the legacy managed Swiftype platform.
- Consolidating content silos from 5+ different web domains into a centralized index.
- Managing a dual-stream data pipeline to handle both legacy web pages and new Headless CMS content simultaneously.
- Delivering high-precision search results across five different languages (English, Italian, Spanish, Korean, and Japanese)

Solution

- Hybrid Ingestion: Deployed Elastic Open Web Crawler for domain scanning and custom Python scripts for Contentstack CMS integration.
- Configuration & Tuning: Established Index Templates, custom Ingest Pipelines, and relevance tuning using Synonyms and Query Rules.
- API Integration: Developed dedicated Autocomplete and Search APIs to provide instant suggestions and full-text results to the frontend.

Solution Delivery

SquareShift migrated a legacy Swiftype instance to Elasticsearch Serverless to eliminate technical constraints and operational costs. It implemented a hybrid ingestion model using the Elastic Open Web Crawler and custom Python scripts to unify search across 5+ domains and five global languages.

To explore the full scope, use the download link below.

Technology Stack

bottom of page