top of page

Language Test

An automated language verification tool designed to validate job seekers' proficiency levels and ensure recruitment accuracy for premium employers.

Jump to prototype
test certification result optimized.png
missandei description image compressed.png

Description

Recruiters frequently face the challenge of "skill inflation," where candidates exaggerate their language proficiency on their profiles, leading to wasted time during the interview process. This project introduces an automated testing suite integrated directly into a recruitment platform’s hiring workflow. By combining traditional multiple-choice assessments for reading and listening with advanced GPT-powered grading for writing, the tool provides a reliable, objective verification of a candidate's true abilities, resulting in a shareable certification and an automatically updated user profile.

Role: UX/UI Designer, UX Researcher

Figma-1-logo.png
dalle chatgpt.png

The goal

To eliminate recruitment inefficiencies by providing a trustworthy, automated verification system that ensures candidates possess the language skills required for a role before they reach the interview stage.

Core functionality

  • Multi-Dimensional Testing: Assesses language proficiency across three key areas: Listening, Reading, and Writing.
     

  • AI-Powered Grading: Utilizes a custom GPT prompt to automatically score written responses based on grammar, vocabulary, structure, and task fulfillment.
     

  • Seamless Workflow Integration: The test is embedded within job screening questions to capture data during the application process without adding manual work for recruiters.
     

  • Automated Profile Validation: Automatically adjusts a user’s profile level to match their test results, preventing candidates from manually overstating their skills.
     

  • Shareable Certifications: Generates a detailed digital certificate showing specific skill breakdowns that users can share on social media or with peers.
     

  • Integrity Controls: Includes timed sections, randomized question variations, and a limited number of attempts (twice every six months) to prevent cheating and ensure score validity.

Research insights

ideation.png

Wireframing

image.png

Style guide

image.png
missandei finished 1.png
missandei finished 2.png

End product

missandei finished 3.png

Final prototype

bottom of page