Description

Nowadays most websites and apps provide a prominent search function by which users can browse through the content repository of a website. This content is usually classified in categories (e.g. event, restaurant, museum) and can contain a diverse set of attributes (e.g. child-friendly). These categories and attributes are generally referred to as tags. Tags can be helpful to offer meaningful search experiences especially when they are combined to filter/present suitable search results (e.g. all restaurants which are classified as child-friendly).  Develop a search algorithm that understands the search intent of a user and identifies one or a combination (intersection) of multiple tags. The algorithm should already consider different tag specifications and spelling mistakes.

Main Tasks

• Design a mechanism to parse and interpret user search queries to identify the underlying intent. The algorithm should analyze keywords, contextual phrases, and synonyms to map queries to relevant tags.
• Implement a system that normalizes tag variations by accounting for synonyms, plurals, and common spelling errors. Ensure it can recognize similar tags across different formats (e.g., “child-friendly” vs. “kid-friendly”).
• Create an intersection logic to combine multiple tags based on user input, ensuring results reflect the most relevant match. Integrate a scoring system to rank search results by relevance to user intent.