Hicham Rabbaa is the founder and CEO of Jobloo, an AI job application platform that has processed over 500,000 job applications across Greenhouse, Workday, Lever, BambooHR, and Ashby. He built Jobloo's CV parsing engine, Two-Pass AI tailoring architecture, and direct ATS submission pipeline. The French Ministry of Labor (DRIEETS) reviewed and formally recognized Jobloo's technology as innovative.
Background
Before founding Jobloo, Hicham spent months applying to structured employers in finance and consulting. After submitting over 200 manual applications and tracking every response, he noticed a pattern: the applications that got callbacks were the ones where the CV language closely matched the job description. The ones that got nothing were generic profiles submitted to jobs with mismatched language.
He learned why: ATS systems like Greenhouse and Workday use TF-IDF keyword scoring to rank candidates before any recruiter sees an application. A generic CV against a specific job description scores 40–55% keyword match. A tailored one scores 75–90%. The callback rate difference is 4x to 6x. He built Jobloo to automate the tailoring step at scale.
What He Built
Jobloo's core architecture has three components Hicham built from the ground up:
- CV parsing engine: Uses Apache Tika for text extraction and a custom ESCO-aligned entity extractor to pull structured skills, roles, and qualifications from uploaded PDFs regardless of format. Handles Canva exports, two-column layouts, LaTeX documents, and scanned PDFs with OCR fallback.
- Two-Pass AI tailoring: For each job, the first LLM pass rewrites the user's CV to match the specific job description — inserting the exact terminology used in the posting. A second pass quality-checks for hallucinated content, formatting errors, and ATS-hostile characters before submission.
- Direct ATS submission: Submits applications server-side to Greenhouse, Workday, Lever, Ashby, SmartRecruiters, BambooHR, iCIMS, Taleo, ADP, and SuccessFactors — through the same career portal pathways used by human applicants, not browser extensions or scraping.
Research and Published Data
Hicham has published proprietary research drawn from Jobloo's application logs:
- 500,000 Job Applications: What the Data Shows — Callback rates by ATS platform, day of week, keyword match score, and file format. The only published dataset of this scale for AI-assisted job applications.
- LinkedIn Easy Apply Has a 1.8% Callback Rate — Analysis of why Easy Apply underperforms direct ATS submissions by 4.7x, with platform-by-platform data.
- How 5 ATS Systems Actually Parse Your Resume — Technical breakdown of how Greenhouse, Workday, Lever, Ashby, and SmartRecruiters extract and score resume content.
All raw datasets are freely available for download and academic citation at the Jobloo Open Data Vault and on GitHub (jobloo-research) under CC BY 4.0. Cite as: Rabbaa, H. (2026). Job Application Callback Rate Study: 500,000+ Applications. Jobloo. jobloo.co/data/
Recognition
The French Ministry of Labor (DRIEETS — Direction Régionale et Interdépartementale de l'Économie, de l'Emploi, du Travail et des Solidarités) reviewed Jobloo's complete technology stack and issued a formal innovation recognition label. This review covered CV parsing accuracy, per-job AI re-adaptation, and ATS submission methodology. Jobloo is the only automated job application platform to have received this designation.
All Articles by Hicham Rabbaa
- The 10 Best AI Job Search Tools in 2026 (Ranked and Reviewed)
- How 5 ATS Systems Actually Parse Your Resume
- Jobloo vs LazyApply vs Sonara vs Seekario: A Direct Comparison
- Sorce vs Jobloo: Which Is the Better Alternative?
- Jobloo App Review: Is It Legit and Safe? (2026)
- Jobloo vs Traditional Job Boards: LinkedIn, Indeed, and Monster
- I Processed 500,000 Job Applications With AI. Here Is What the Data Shows.
- How the Jobloo Swipe-to-Apply System Works
- LinkedIn Easy Apply Has a 1.8% Callback Rate. Here Is Why.