INTERNSHIP - AI driven keywords and concept detection for an automated grading system

Engineer – Software

Izegem, Belgium

Televic develops, manufactures and installs top end high-tech communication systems for specific niche markets. A financially independent and stable group, Televic is divided into divisions that each focus on their specific market: 

Televic Rail: passenger information systems and on-board control systems for trains
Televic Healthcare: communication systems for healthcare
Televic Conference: conference systems for large venues 
Televic Education: multimedia and e-learning solutions for staff training and educational institutions

Televic creates added value for its customers by developing custom-made solutions and by continuously innovating at the cutting edge of technology. With headquarters in Belgium and offices and plants across Europe, Asia and the US, Televic employs over 1000 people worldwide.

About Televic Education

Televic Education is a world leader in the development of research-based solutions that help solve training, quality, certification, accreditation, permanent evaluation and collaboration challenges in many different industries. The combination of our in-depth knowledge of specific sectors with scientific research has enabled us to create innovative solutions and to build long-term relationships with our clients that range from governments, educational institutions, hospitals as well as corporates.


assessmentQ is a digital exam platform developed and maintained by Televic Education. 250+ customers are using assessmentQ for organizing online and digital (practice) exams.

This thesis aims to address the challenge of automated grading of open-ended questions, a task currently unresolved in many e-learning platforms. Presently, our platform utilizes a manual keyword-based grading system, which, while effective, demands significant time and effort to establish keyword rules. The primary objective of this thesis is to devise an innovative solution capable of autonomously identifying and associating keywords to form coherent concepts, thereby streamlining the grading process for open-ended questions and enhancing the overall efficiency of the automated grading system.

The end goal of this thesis is the development of a proof-of-concept system that can:

  • Automatically detect keywords
  • Automatically link certain keywords together (to form a concept) 
  • Outputs a format that can be adapted in our platform assessment
  • (bonus) outputs some form of reasoning behind its decisions


  • Level: Academic Master, Bachelor, Master
  • Specialty: AI / Machine Learning, IT, Software
  • Type of work: Research: 30% Implem.: 40% Experim.: 30%
  • Location: Televic, University
  • Type of activities: Experimenting, Implementation, Literature study, Programming
  • Number of students: 1 or 2


Personal information
How would you like us to contact you?
One file only. 15 MB limit. Allowed types: pdf, docx, doc.
One file only. 15 MB limit. Allowed types: pdf, docs, doc.
Extra information
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.