Hello Haplos: Building the AI-Powered Job Search Tool We Wished Existed
Introducing Haplos — how we built an AI resume optimization tool that actually helps job seekers land more interviews, and what we learned along the way.
Hello Haplos: Building the AI-Powered Job Search Tool We Wished Existed
Job searching is exhausting. You spend hours tailoring your resume for each application, trying to guess what keywords the ATS (Applicant Tracking System) is scanning for, and wondering why you never hear back from jobs you're clearly qualified for.
We built Haplos because we've been there — and we thought there had to be a better way.
The Problem with Modern Job Searching
Most ATS systems are dumb filters. They scan resumes for keyword matches, score them on a rubric, and only pass a small fraction to human eyes. The frustrating reality? A great candidate with a resume that doesn't match the job description's exact phrasing gets filtered out before anyone even reads it.
This creates a perverse dynamic: you need to optimize your resume for robots before it ever reaches a human. But manually doing that for every job you apply to is:
- Time-consuming (30–60 minutes per tailored application)
- Demoralizing (it feels less like showcasing yourself and more like keyword stuffing)
- Error-prone (easy to miss important terms buried in a 500-word job description)
We wanted to fix this.
What Haplos Does
Haplos takes your master resume and a job description (or a job posting URL), and generates a tailored version optimized for that specific role. Here's what happens under the hood:
// Simplified view of the tailoring pipeline
async function tailorResume(resume: string, jobDescription: string, intensity: 'slight' | 'medium' | 'high') {
const keywords = await extractKeywords(jobDescription)
const gaps = identifyGaps(resume, keywords)
const tailored = await rewriteWithGaps(resume, gaps, intensity)
return tailored
}The intensity parameter is key. Not every job warrants the same level of rewriting:
- Slight: Minor keyword additions, preserves your original voice
- Medium: Moderate rewriting that balances authenticity with optimization
- High: Significant rewriting for maximum ATS score — useful when you're making a career pivot
What We Learned Building It
1. Prompt engineering is iterative
Getting the AI to rewrite resumes in a way that feels authentic rather than robotic took dozens of iterations. The key insight: you need to give the model a clear target (the job description) AND clear constraints (don't fabricate experience, preserve the candidate's voice, maintain truthfulness).
2. Users care about control
Early testers wanted to see what changed. We added a diff view that highlights modifications so you always know exactly what Haplos touched.
3. ATS optimization is a moving target
Different companies use different ATS systems, and they update their algorithms. We're building a feedback loop using anonymized application outcome data to continuously improve our tailoring logic.
What's Coming Next
We're just getting started. The roadmap includes:
- Cover letter generation — tailored to the same job posting, in your voice
- Interview prep — AI-generated practice questions based on the job description and your experience
- Job match scoring — before you apply, understand how well your background fits the role
- Application tracking — a lightweight CRM for your job search pipeline
Try It Free
Haplos gives you 3 free tailored resumes with no credit card required. We think once you see the difference a properly optimized resume makes, you'll want to use it for every application.
The job market is tough. Your resume shouldn't be the reason you don't get an interview.