5x faster job search
Improving Handshake’s core experience for millions of early-career jobseekers
Job search is Handshake’s most-used experience and the primary path to application conversion. In late 2024, I led the design improvements effort: migrating it to Handshake’s design system, simplifying the filter interaction, and making it responsive for the first time. The result was a faster to use job search with a better mobile-web experience that led to higher job views and applications.
Role: Design lead
Scope: UX strategy and design
Timeline: 2 months


Why job search needed upgrades
Job search is Handshake’s highest-traffic surface and directly tied to application conversion.
Before this project:
- Filters were confusing and inefficient
- Experience was slow to load
- 35% of traffic was on mobile web but the layout was not responsive
- Visually inconsistent with the rest of the product
- Built on an outdated design system


As Handshake expanded access beyond school-affiliated students, significantly more users would enter this flow. The experience needed to feel modern, fast, and cohesive before that shift.


A foundation-first approach
Improvements needed to preserve product stability and metrics. I approached the project with three guiding principles.
1. Aligning job search with the design system
Job search had diverged from the rest of the product. The first priority was aligning it with Handshake’s design system. This would offer free benefits including responsiveness and accessibility, raising the baseline quality of this experience.


2. Improving filters without risking conversion
Because job search directly impacts application conversion, PMs were cautious about introducing behavioral changes, particularly around filtering.
The legacy "All filters" interaction opened a dense modal containing mixed control types that behaved inconsistently. Some selections disappeared after applying filters, and every selection unnecessarily refreshed the page behind it.
Rather than redesigning filters, I focused on making the interaction clearer and more consistent. I restructured the modal, leaning into our proven design system patterns to provide:
- Clear grouping and hierarchy, relying on simple, universally recognized controls
- Consistent interaction patterns across controls
- Predictable selection states
This also eliminated unnecessary page reloads, improving performance.
3. Using architecture to improve speed and responsiveness
Migrating job search to our design system's 2-pane responsive layout simplified the architecture. The layout component offered filtering and search logic which we used to unlock a faster experience.
The 2-pane layout from the design system also added responsive support. Roughly 35% of traffic came from mobile web and this allowed the product to scale across screen sizes without creating a separate mobile experience.
Additionally, designing skeleton loaders helped improve perceived speed, reducing the load times that customers complained about.
The new experience


The result was a job search that looked and behaved like the rest of Handshake: consistent components, faster load times, responsive layout, and filters that worked predictably. For users, that meant spending less time fighting the UI and more time finding and applying to jobs.


Results
Load times dropped from 5s to 1s - a 5× improvement. Job views increased 6% and applications increased 5%, with overall traffic up 2%. The responsive layout extended the experience to the 35% of users who had previously been on an unsupported mobile layout.


Lessons from updating a high-traffic experience
- Clarity beats novelty when conversion is the metric. I restructured the filter modal without changing what it does, opting for more straightforward, familiar interactions.
- Design system adoption as a product lever. Aligning job search with the design system unlocked responsive behavior, improved accessibility compliance and faster loading. All this helped move key metrics meaningfully without rethinking the experience.
- Risk aversion forced better scope. "No behavioral changes" felt like a blocker at first. In practice it focused the work toward what actually mattered: clarity and system alignment, not reinvention.

