How We Built an AI-Powered Job Application Tracker

Introduction

Jobber started as an internal tool. Engineers at fluxLab.dev were tracking job applications in spreadsheets — losing track of which resume went where, which companies they'd heard back from, and which interviews were next. We built Jobber to solve our own problem, then realized every job seeker has the same pain.

The Problem

Job searching is chaotic. A typical active job seeker:

Applies to 20-50 positions across multiple job boards
Has 2-3 resume versions for different role types
Tracks applications in a spreadsheet that becomes outdated within a week
Misses follow-up deadlines because there's no reminder system
Has no idea which resume version or application source performs best

The Solution: Jobber

Jobber is a personal job search management platform. It handles the entire pipeline from finding a job to tracking the outcome:

Job Import

Users save jobs from any website using a Chrome extension. When they click "Save to Jobber," the extension captures the page content and sends it to our backend.

The backend uses a three-layer parsing approach:

JSON-LD extraction — structured data from LinkedIn, Indeed, Glassdoor (45% of imports)
DOM pattern matching — site-specific selectors for known job boards (30%)
Claude AI fallback — LLM extracts job details from raw page text (19%)

This layered approach achieves a 94% success rate while keeping AI costs minimal.

Application Tracking

Applications are organized in both list and Kanban board views. Users can:

Track status: Applied → Screening → Interview → Offer → Rejected
Add notes and comments at each stage
Set custom stage templates that match their interview process
Filter by company, status, or date range

AI Resume Builder

Jobber includes a full resume builder with AI assistance:

Section management: Work experience, education, skills, projects, certifications
AI summary generation: Creates a professional summary from work history
Bullet point suggestions: Generates achievement-focused bullet points
ATS compatibility check: Ensures resume format works with applicant tracking systems

Resume-to-Job Match Score

Users can check how well their resume matches a specific job posting. The AI analyzes:

Skill overlap (technical and soft skills)
Experience level alignment
Missing keywords
Category-by-category breakdown

This helps users decide whether to apply or tailor their resume first.

Cover Letter Generation

For each application, Jobber generates a tailored cover letter based on:

The job description and requirements
The user's resume content
Company information

Users can edit, improve, or regenerate cover letters with AI assistance.

Analytics

The analytics dashboard provides insights that spreadsheets can't:

Application funnel: Conversion rates through each stage
Time in stages: Average days spent waiting at each step
Resume effectiveness: Which resume versions get the best response rates
Source performance: Which job boards produce the most interviews
Response rate: Overall and per-company callback rates

Tech Stack

Frontend: Next.js, TypeScript, Tailwind CSS
Backend: Go with chi router, pgx for PostgreSQL
Database: PostgreSQL + Redis
AI: Claude API (Haiku for parsing, Sonnet for generation)
Infrastructure: Docker, Hetzner Cloud, Caddy
Payments: Paddle for subscription billing

What's Next

We're working on Google Calendar integration for scheduling interviews directly from Jobber, reminder notifications for follow-ups, and a tagging system for organizing applications by priority.

Conclusion

Jobber proves that AI features don't need to be the entire product — they enhance a solid workflow. The core value is organization and visibility. AI makes it faster and smarter. If you're actively job searching, give Jobber a try at jobber.fluxlab.dev.