Redesigning the Job Search Experience for International Students
Platform
Web Application
Model
B2C
Category
Job Search
Year
2024 - 2025
This project was a 6-month collaboration with the job-hunting platform focused on international students from Eth Tech company. The team aimed to improve engagement and usability through a full UX and UI redesign. The platform targets a specific audience with unique constraints during job hunting, such as visa sponsorship requirements and limited timeline.
My Role
I served as the sole product designer, leading the entire Double Diamond process, including user research, flow definition, wireframing, UI design, usability testing, prototyping, and iteration.
Collaborations
Product manager, backend engineers, marketing analysts.
Tools
Figma, Miro, Google Form, Google Sheet, Protopie
Timeframe
Six months, structured around the Double Diamond framework.
-18%
Daily Active Users
-15%
Conversion Rate
+10%
Drop-off Rates
According to the internal data from Product Manager, from June to September it showed critical drops across key metrics:：Daily Active Users decreased by 18%； Conversion rate from search to apply dropped by 15%；Abandonment rates rose steadily； Retention and engagement continued to decline.
Developed using Bootstrap without proper design structure, the previous website struggled to resonate with users. And the team believed that the declining metrics were primarily due to the site’s outdated look and lack of a youthful, engaging tone. Based on this assumption, I was brought in to redesign and upgrade the overall design language
I followed the Double Diamond framework to guide the entire project. This process helped ensure that every design decision was grounded in user research and systematically refined.
Discover: Identified problems through surveys, interviews, and competitor analysis.
Define: Synthesized insights into clear pain points, personas, and design goals.
Develop: Explored ideas through sketches, wireframes, and early testing.
Deliver: Refined high-fidelity designs and validated features with usability tests.
Competitor Analysis
I did researches on four job platforms—two with broad job coverage and two more student-focused—to understand current market offerings for international job seekers. Using SWOT analysis, I evaluated each platform’s strengths, weaknesses, threats, and opportunities. While some provide comprehensive listings, visa tags, or STEM-focused roles, most lack consistent user flows, clear visual hierarchy, and tailored features for international students. These gaps create an opportunity to combine targeted H-1B filters, student-first tools, and modern, trust-building design to improve usability, engagement, and application conversion.
15 Interviews
92% USCIS
97% External Apply
80% Local Track
90% Poor Visa insight
89% Inefficiency Flow
Target Audience
Eth Tech job posting website has around 10,000 active users, with 90% international students. Most are in STEM fields or aiming to transition into STEM careers. 85% are undergraduates and 13% are graduate students.
Primary: International STEM undergraduates seeking full-time jobs after graduation
Secondary: Non-STEM students looking to transition into STEM roles
1M Existing Users
85% Undergraduate
13% Graduate
90%
International STEM Student
Interview Insights
In 25 user interviews, 96% were international students and 88% were recent graduates or about to graduate. Participants valued our website for its immigration information, which general job platforms like LinkedIn and Indeed do not provide.
However, the drop-off rate while using our site was notably high. When asked where they went after leaving the platform, 92% visited USCIS or similar sites for company sponsorship history, 97% applied through external portals, and 80% tracked applications in Excel, Google Sheets, or Notion. The main reasons were the lack of a direct application portal and a job tracker that felt incomplete and hard to customize.
Affinity Mapping
To make sense of the survey and interview findings, I worked with the UXR to organize all data into an affinity map. Insights were grouped into three key categories: user pain points, user needs, and market opportunities. This visual clustering helped reveal patterns across feedback, such as recurring frustrations with visa information, inefficiencies in job search, and unmet needs in application tracking. The map served as a shared reference for cross-functional discussions, ensuring everyone aligned on the core problems and opportunities before moving into ideation.
Personas
To make sense of the survey and interview findings, I worked with the UXR to organize all data into an affinity map. Insights were grouped into three key categories: user pain points, user needs, and market opportunities. This visual clustering helped reveal patterns across feedback, such as recurring frustrations with visa information, inefficiencies in job search, and unmet needs in application tracking. The map served as a shared reference for cross-functional discussions, ensuring everyone aligned on the core problems and opportunities before moving into ideation.
With the research insights consolidated, I collaborated with the product team to recalibrate our design goals. We shifted from the initial assumption of a purely visual redesign to addressing the broader need:
“Improving the end-to-end job-seeking experience”
After aligning on the goal, we established clear design priorities that would guide the upcoming ideation phase.
Addressing What Matters Most
Based on research insights, I brainstormed and organized potential feature solutions to address key user pain points and market opportunities. These ideas were evaluated against user impact, feasibility, and alignment with business goals. The initial focus was on improving visa-related information, streamlining job search efficiency, reducing the learning curve for career switchers, and simplifying application tracking.
Feature Prioritization Using Matrix
To evaluate and prioritize potential solutions, I plotted each idea on a Feature Priority Matrix based on two dimensions: user value and implementation effort. This approach allowed the team to visualize trade-offs, identify quick wins, and plan phased delivery.
Mapping User Flow
With the top-priority features identified, the next step was to map how they would fit into the user’s real journey. I translated each idea into actionable steps within the job search process, ensuring they worked together as a cohesive experience. The red-framed sections highlight features prioritized in the previous matrix exercise, now placed within the full flow to visualize their role in the journey. By mapping the end-to-end flow, the team could anticipate user decisions, streamline key interactions, and ensure every high-value feature was seamlessly integrated.
Skecth
After mapping the user flow, I sketched four key features to visualize how users move from search to application. These sketches then guided the wireframes, where the designs are shown in clearer detail.
Wireframe
Design 1
On the homepage job description section, a new Company H1B Insights panel and an AI Summary panel are added. In the job description text, technical terms can be hovered to display tooltips with explanations and a direct Google link.
Design 2
Introduces an internal application flow. By clicking the Apply button, a modal appears where users can upload their resume and supporting documents, allowing them to submit directly without being redirected to an external site.
Design 3
An Application History management page designed in card layout. Each application is displayed as a card, with a new Sync to Tracker feature allowing users to select (via checkboxes) which applications they want to sync.
Design 4
After selecting applications, users are prompted with a modal to choose a sync destination (external platform) or export as a local file (Excel, CSV, PDF).
Design 1
Design 2
Design 3
Design 4
Redesigned Home Page
To optimize the use of white space on the job description page, the Job Match score, AI Summary, and Company H1B Insights panels were positioned on the right side as supplementary content, ensuring that the job description itself remains uncluttered.
Icons were introduced in the text to highlight key information for faster scanning and comprehension. The AI Summary was redesigned from a long paragraph into categorized tags, making the content clearer and more user-friendly. The Job Match panel was enhanced with explanatory cues so users can quickly understand how each requirement contributes to the overall score.
Internal Apply Feature
Building on the wireframe design, the internal apply flow was refined to improve clarity and usability. A step indicator (Step 1 / Step 2) was added below the progress bar, giving users a clear sense of how many steps are required to complete the process. In addition, an edit icon was placed next to the Resume edit button, making it more visually prominent and allowing users to quickly identify and access the edit action.
Sync to Tracker
Expanding on the wireframe design, each application card now displays the submission date, giving users clearer context for their application history. A pagination control was added at the bottom of the panel to indicate the total number of pages, helping users better navigate large histories. The “Sync to Tracker” action was redesigned from plain text into a button, making the feature more prominent and visually accessible.
Select Tracker Platform
Based on the wireframe design, each platform option was enhanced with its corresponding icon, allowing users to quickly recognize and select their preferred sync destination. The processing state was improved with a motion illustration and an animated progress bar to convey a clear sense of ongoing activity. Finally, the success state now features a checkmark illustration, giving users a strong visual confirmation that their sync has been completed successfully.
On Maze, I tested the existing high-fidelity prototype by designing tasks for users to complete. These tasks were created to let participants try out the newly added features.
During the test, I observed how long users spent on each task, where they paused or struggled, and gathered their direct feedback. I then summarized these insights and applied them to guide the following design iterations. (The figure below shows screenshots from Maze results, including user feedback on the entire high fidelity design flow; the heatmaps of clicks on the glossary feature, as well as data on page duration time and mismatches.)
Based on usability testing and design reflection, I refined several key features to better align with user needs and improve overall clarity.
Surfacing H-1B Insights at First Glance
Users felt it was inefficient to click each job card to access H-1B insights, the information they valued most. I extracted the two most critical data points—approval rate and sponsor ranking—and displayed them directly on the job card, allowing users to scan company eligibility faster without extra clicks.
Making Glossary Use More Active
Testing showed the glossary feature was underused because it felt too passive. To address this, I kept term highlighting but added a floating button indicating how many technical terms were annotated on the page. When users hover, the terms gently flash, and the definition pop-up now has a gradient border to feel more intelligent and engaging.
Enhancing Feature Visibility and Clarity
Additional refinements included adding theme-colored top bars to new feature cards for stronger visibility, and replacing the close “X” icon with “+” and “–” symbols to clearly convey expand/collapse behavior.
Below is a collage of the final design screens. To explore the interactive prototype, please click this link
+ 17.2%
Conversion Rate
+ 9.8%
User Satisfaction
3 MVP
Launched
The redesign delivered measurable improvements across both business and user experience metrics, which validated the effectiveness of a user-centered, iterative design process and demonstrated tangible value to both end users and the business:
• Improved conversion rate by 17.2% after introducing a streamlined job search-to-apply flow.
• Increased user satisfaction by 9.8% through consistent design system updates and polished visual language.
• Successfully launched 3 MVP features from 0 → 1 (AI job insights, application tracker, glossary), ensuring smooth product reviews and adoption.
• Enhanced design workflow efficiency by creating and maintaining a Figma design system, covering iconography, grids, and reusable components.