CareerCraft

Inspiration

Job hunting is stressful enough—but tailoring resumes, writing cover letters, and keeping track of each application? That’s where things fall apart. I built CareerCraft to solve that problem.

I was inspired by how many talented people I knew—myself included—who either sent out generic applications due to time constraints or burned out trying to personalize every submission. With the rise of accessible AI tools like GPT-3.5, I saw an opportunity to simplify and humanize this process with a fast, smart assistant that helps job seekers focus on their strengths, not their formatting.

What it does

CareerCraft is a web-based AI-powered job application assistant that automatically generates tailored documents based on the user’s profile and the job they’re applying for. It offers:

• Profile management: Users can enter personal info, work experience, education, and skills.
• Job input: Paste a job description or submit a job URL (LinkedIn, Indeed, etc.).
• AI parsing: Extracts role-specific keywords and requirements from job posts.
• Document generation: Tailored resumes and personalized cover letters—ready to preview and download.
• Smart prompting: Designed to avoid hallucinations or generic filler, resulting in content that genuinely fits the user and the job.

The result? High-quality applications generated in minutes—with the user always in control.

How we built it

CareerCraft was developed as a frontend-only MVP using modern web tools:

• Frontend: React 18 with TypeScript and the Vite build tool
• Styling: Tailwind CSS with a custom design system
  
  • Primary: #2563EB
    
  • Secondary: #1E40AF
• Icons: Lucide React
• AI Integration: OpenAI GPT-3.5-turbo via direct fetch calls
• URL Parsing: Custom service using allorigins.win to fetch and parse job descriptions
• Data Storage: Browser localStorage for session persistence
• Code Quality: ESLint

The development was driven through micro-steps using Claude for prompt strategy and Bolt.new for implementation. Every feature was tested in isolation with a focus on context retention, usability, and graceful failure recovery.

Challenges we ran into

• Infinite loading bug: During the final hours of submission, the Devpost platform froze at the video upload step, preventing successful entry submission despite the project being ready.
• Job parsing reliability: Each job site had different HTML structures, so building a parser that worked across LinkedIn, Indeed, and custom pages required fallback mechanisms.
• Prompt engineering: GPT-3.5-turbo occasionally hallucinated experiences or skills. We had to build a strict prompting system to constrain the AI’s imagination and enforce accuracy.
• Form complexity: Creating an intuitive, multi-step profile form that supports dynamic fields (e.g., multiple work experiences) without confusing the user was a major UX task.

Accomplishments that we're proud of

• Built a fully functional AI job application assistant with no backend.
• Implemented dual job input modes (URL + paste), along with AI-powered job description parsing.
• Designed and enforced anti-hallucination prompts for more truthful AI outputs.
• Created a responsive, professional UI with real-time validation, smart feedback, and user-centered flows.
• Enabled users to preview and download tailored documents instantly as PDFs ready for submission.

What we learned

This project was an incredible lesson in balancing AI automation with human authenticity. I learned how vital prompt engineering is when the goal is not just to generate content—but to represent someone’s identity through that content.

On the technical side, I gained deep experience with browser-first architecture, fallback logic, CORS workarounds, local persistence strategies, and building for error resilience in AI workflows.

Most importantly, I learned how powerful it is to ship small, test constantly, and build with empathy for users under pressure.

What's next for CareerCraft

Now that the core MVP is complete, the next steps are focused on product polish and scale:

• 🎨 Template System: Let users choose between different resume/cover letter styles
• 🛡️ Backend Integration: Secure API key handling and user accounts via Supabase
• 📈 Application Tracker: Help users manage where they've applied and view AI usage history
• 📊 Analytics Dashboard: Visual insights into what roles users are applying to
• 📱 Mobile App: Bring the experience to phones for job seekers on the go
• 💸 Monetization Strategy:
  • Freemium Model: Offer core features for free, with premium perks like additional templates, PDF exports, and usage credits behind a subscription.
  • Pay-per-use: For occasional users, enable one-time payments to generate documents or unlock specific features.
  • Team/Agency Plans: Provide volume-based pricing for career coaches, bootcamps, or HR teams to use with their clients or cohorts.
  • Affiliate Partnerships: Integrate with job boards, resume feedback tools, or career services to create value-driven affiliate channels.

CareerCraft isn’t just a tool—it’s a mission to make applying for jobs less painful and more personal, using the best of what AI has to offer.

Built With

• bolt.new
• claude
• netlify
• openai
• supabase