Job Hunt AI Buddy | UX Page Redesign
MY ROLE: Product Designer & Developer
TOOLS: Figma AI, Claude Cowork & Code, GitHub
DURATION: 20+ hours
Job Apps CRM AI Dashboard with Apply Assist
Problem
Job seekers today are showing higher stress levels — driven by AI-related layoffs, ghost job postings, and the daily grind of reapplying on repeat. On top of all this, job seekers have to track their jobs applied through various methods from spreadsheets to paid tools. On the other side, company recruiters have access to enterprise-grade AI screening tools built to find exactly the candidates they want for their UX needs.
Solution
FitPilot helps level that playing field. It scans for relevant postings, helps applicants organize jobs by user actions, has AI assistance which tailors resumes and cover letters to each one, and flags likely ghost jobs before time gets wasted on them. The job seeker still reviews the AI-assisted results and applies manually, keeping it genuine enough to avoid getting flagged by employer screening. It becomes an end-to-end job application dashboard experience.
Product Demo
Role
I recognized how repetitive and exhausting job applications had become — endless tailoring, forms, and ghost postings. So I built a tool that scans jobs, organizes jobs based on user actions, tailors resumes and cover letters, and flags shady listings. I designed it like a recruiter’s ATS dashboard, but for job seekers screening postings instead of candidates. Then I guided Claude Code to help build it as a former developer. I deployed it as a public open-source tool on GitHub for other job seekers facing the same challenges.
What Others Do Well
Our team did product research on direct and tangential competitors. Mine included Asana, Monday, and ClickUp of their pricing pages.
We used Claude CoWork to analyze market product needs met from competitors.
Jobsight
- AI copilot actively finds and scores relevant postings — no manual searching needed.
- 1-click autofill extension makes applying nearly instant once a job is chosen.
- Runs on a proprietary cloud model, so nothing stays local.
- No way to check whether a posting or company is actually legitimate.
Careerflow AI
- Focuses on resume scoring and improvement suggestions rather than automating anything.
- Never submits or tailors on the user’s behalf — lowest-risk of the three.
- Doesn’t scan for jobs or tailor materials per posting.
- Best suited as a companion tool, not a full application workflow.
Applykit
- Tailors resumes with a local LLM (Ollama), so nothing leaves the user’s machine.
- Fully open-source and free — closest philosophy match to this project.
- Doesn’t scan for postings or check legitimacy.
- Autofill and final-submit behavior aren’t clearly documented.
Survey & Research
We used Claude CoWork to analyze user research for affinity map and other findings.
Based on user needs, we did more research on aspects of the job application process the users wanted. We found various products filled some needs, but not others.
| GhostBust | Teal | Simplify | |
| Find Jobs | ❌ | ❌ | ✅ |
| Job Application Tracker | ❌ | ✅ | ❌ * |
| Resume Tailoring | ❌ | ❌ | ✅ |
| Job Apply Autofill | ❌ | ❌ | ✅ |
| Ghost Job Verification | ✅ | ❌ | ❌ |
| Company Verification | ❌ | ❌ | ❌ |
Survey
We put together a quick Google survey asking what frustrations users had (50 users). These were the most 2 relevant user questions from the data.
Affinity Map
Repetitive busywork is pain
forms, tailoring, and cover letters were the top three “most tedious” tasks, and the app’s core loop targets exactly that.
Generic AI output kills trust
“felt generic” was the #1 AI complaint, so generation stays a deliberate, reviewed step, not a black box.
Ghost jobs are a big fear
nearly 1 in 3 seekers have applied into a fake posting, which is why legitimacy checking made the cut.
Ideation
As a user, I want a tool that finds real job postings and tailors my resume for them.
Produce Values
Taking into account the user needs from the survey and the affinity, we chose to build our product on trying to capture all the users’ pain points into one product from end to end.
I. Jobs matched to user needs with custom user flagging
II. AI tailored resumes with full manual approval for user actions
III. Job Application Tracker & Organizer with user customization
Wireframe Brainstorming
I looked at many dashboards, using salesforce as a UI example. I then looked at recruiter dashboards as inspiration.
Usability Testing – Wireframe
Desktop Design
Claude CoWork was used to synthesize granular data findings from usability testing.
Wireframe Usability Test Results
1. Generate a tailored resume.
2. Apply for a job with AI-assisted tailored resume.
3. After applying, mark that job as “applied”.
- ✅ All of our users were able to apply for a job and marked it as applied (5/5 user)
- ❌ The users voiced the steps were oversimplified for the usability test
(due to technical limitations, this was simplified. You could have to go into the Claude AI to get the files and apply manually for true use)
Pros
- users able to track jobs applied
- users able to use tailored resumes
- user able to previous jobs (archived)
- users able to easily navigate to previous jobs if they receive callback from company from interview vs other manual tracking system
Cons
- user not exposed complex claude console/local desktop file/manual application proces
- user not exposed to having to start up code to have webapp work locally (setup required)
- user not exposed to hardcoded future roadmap functions (job type, job search, time duration of job search, job location/distance radius, etc)
- no search function in initial version – manual browser word search
High Fidelity Testing Part I
We used the same testing script as the lo-fidelity test. We added some functionality buttons and some UI improvements.
1. Generate a tailored resume.
2. Apply for a job with AI-assisted tailored resume.
3. After applying, mark that job as “applied”.
High Fidelity Part I Usability Test Results
- ✅ All of our users were able to apply for a job and marked it as applied (5/5 user)
- ❌ Users did want more functionality per usual job site functions built out
- ❌ Users did want a more responsive UX process with UI elements to confirm their actions to be reflected
- ❌ Users did report vagueness in production function despite understanding product – too barebones
Claude CoWork was used to synthesize granular data findings from usability testing.
Post High Fidelity Part I User Test Changes
The users indicated a lack of UX features and UI polish that preexists in usual sites like LinkedIn for example. I introduced a number of these features users would expect from a current feature product.
Job Customization
Users can now customize the jobs they want ✅ with job title or companies for the scanner automatically searches for.
User Customization
Users can upload their resume ✅ for the AI as a baseline to customize for jobs they want to apply for.
Applied and Hidden Tabs
Users can see jobs hidden or applied ✅ with their own tab for easy organization vs a manual method.
Active Search Filters and Job Options
Users can filter their job results ✅ just like a regular job search site with a second level of filtering.
Users on job cards can favorite, hide. flag, and delete ✅ jobs and let the system know how to handle certain companies or job titles in the future.
Pre-Verified Jobs
Users can see jobs that are pre-verified ✅ per a established company name like Meta, Microsoft, etc.
Users are also able to see the job listing, job source (LinkedIn, etc), date, and other common details easily.
Unverified Jobs
Users cans see jobs that can be user verified ✅ upon user verification of smaller companies maually and would be verified in the future.
Users are also able to see the job listing, job source (LinkedIn, etc), date, and other common details easily.
start here
High Fidelity Testing Part II
1. Set up initial job scan settings.
2. Apply for a job with AI assistance.
3. Change job and user settings.
- ✅ All of our users were able to apply for a job and marked it as applied (5/5 user)
- ❌ Users did want more functionality per usual job site functions built out
- ❌ Users did want a more responsive UX process with UI elements to confirm their actions to be reflected
- ❌ Users did report vagueness in production function despite understanding product – too barebones
Claude CoWork was used to synthesize granular data findings from usability testing.
Final Design
Site Improvements
❌ The existing site has confusing iconography and difficult typography to read
✅ The plans module looks modern and easy to navigate with relevant information.
❌ The existing site has seemingly random CTAs below the plans for UX content organization.
✅ The Add Ons Module is a modern carousel and easily scannable.
✅ Additional CTA callout at the bottom for user conversion.
❌ The Plans details gives too much irrelevant information and has poor visual design.
✅ The Plan Details Module is a modern collapsible accordion with key content.
Here is the final deliverable we gave to the client. We incorporated their request for final content and the Kanban design system.
Click to enlarge images.
Desktop
Mobile
Desktop and Mobile Data Accordions
Client Feedback
Initially, the client was hesitant about conducting UX testing, but with the support of our combined UX testing results, they were very satisfied with the design. We maintained the existing design system’s look of the page while giving it a visually appealing modern feeling that was user-friendly while accessible.
Live Product
The client made some minor changes and added some functionality updates (language selector) and the page is live. Subject to change by client since project went live circa 2022.
Outcome
Increased prodcut signup by 50%
Our design team does not have the exact web analytics, but analytics from App Store downloads and product user traffic reviews strongly indicate conversion increased by 50% and the redesigned page has been very well received with user feedback.