Useme - Job Board search refactor

Business problem

The job search needed functional improvements (missing search options) and visual cleanup (simplification, view hierarchy). The database, under heavy load, required read operation optimization to minimize page load times. Additionally, the SEO tag database was too small, limiting Google visibility.

Approach and solution

I co-developed the search rebuild with full filtering by phrase, categories, subcategories, and skill tags. I denormalized key tables, which reduced database load and drastically accelerated read speeds. We expanded the SEO tag system with bulk CSV import/export. Everything was deployed under feature flags for safe rollouts.

Delivery scope

• Search panel with phrase input and Search button - static search triggered on click.
• Sidebar with category list: clicking reloads the page and shows subcategories. Categories matching the phrase highlighted in black, others grayed out.
• Skills (tag) search with suggestions - narrowing results by tags attached to jobs and present in listing text.
• Data model denormalization - planning and implementing a strategy of duplicating key read data to reduce JOINs and speed up search queries.
• SEO tag system: bulk CSV upload to admin panel, tag database export, improved tag suggestion in the skills input.
• Optimization of Find Freelancer and Find Job views.
• Deployments under feature flags - splitting large changes into smaller steps with rollback capability at any point.

Business impact

• Simplified and more intuitive job and freelancer search.
• Increased page responsiveness and faster read operations due to data denormalization and reduced database load.
• Larger SEO tag database through CSV import - better Google visibility.
• Safe deployments without disrupting application availability.
• Search result narrowing by tags attached to jobs and present in listing content.