Loading project details.
Cek CV
A CV checker that simulates how applicant tracking systems parse resumes and provides structured AI-driven insights for improvement.
Context & problem
Many job seekers do not understand how ATS systems interpret resume text structure. The goal was to make parsing behavior transparent, focusing on how text sections and headings are read by ATS rather than visual layout, while keeping user data handling minimal and privacy-conscious.
Outcome
Resume sections are shown with AI feedback; uploaded data is not kept after processing.
Engineering decisions
Built a rule-based parsing engine to simulate ATS text interpretation, prioritizing deterministic and explainable output. Layered AI-powered insight generation on top of structured parsing to provide contextual improvement suggestions, while enforcing usage limits to control API cost and keeping resume data processing temporary (no persistent storage).
My contribution
Designed structured resume parsing visualization to make each parsed section explicitly traceable and ATS-readable.
Implemented explicit rule-based parsing logic to simulate how ATS systems interpret text sections and headings.
Integrated AI-powered insight generation with controlled daily usage limits to provide contextual improvement suggestions.
Structured prompt layering to ensure AI-generated insights remained aligned with deterministic parsing results.
Ensured resume data is processed temporarily without persistent storage to maintain user privacy.
Designed feedback outputs to clearly explain how resume sections are interpreted and where improvements are needed.
Trade-offs
- Chose transparency over complexity in parsing logic.
- Avoided overfitting to specific ATS vendor behavior.