Michal Kabat (Information Engineering with Network Management) and Roman Velic (MSc Information Engineering) have been working as Systems Developers with Tycom, an Aberdeen-based software services company, while studying for their Masters courses at RGU. This is not the only thing they have in common. They both graduated with a first class Hons degree in Computing for Internet and Multimedia from RGU in 2012 and both undertook a Erasmus summer placement in the summer between year two and three of their degree.
To continue with this trend, they decided to split a software project, assigned to them by Tycom, into two parts to tackle as part of their MSc project. The problem at hand was to create an Intelligent Decision Support System for Competence Assurance with a case study in the oil and gas industry.
A simplified description of this problem is as follows:
A company has various job positions. Each job position requires an employee to be competent at various skills. To prove that an employee is competent in a skill, some evidence needs to be presented. The evidence can be in the form of answers to some questions, scanned certificates, observation while the employee is undertaking the job (recorded as text or audio/video). The evidence is then assessed by an assessor. Furthermore, in order to maintain a consistent level of standard (competence) across the personnel, a verifier is used to confirm whether that standard has been achieved. This repetitive task is both time consuming and human resources intensive and there was a need for some intelligent automation to be put in place.
Michal was in charge of developing the part of the system that automatically identified the set of skills related to the evidence submitted by an employee. This was achieved through media-to-text conversion allowing employees to submit evidence is various types of formats. The text is then analysed using techniques pertaining to data mining and natural language processing is order to find the best match of skills to the submitted evidence.
Roman’s part of the system was then to look into the evidence to assist the assessors in determining the level of competence of the presented evidence. It also allows verifiers to agree or disagree with the assessor’s judgement. The system uses techniques of data mining, natural language processing and case-based reasoning to continuously learn from positive as well as negative episodes of assessing and verifying.
Put together, the two parts work in harmony to provide a novel and effective solution to assisting in managing competency assurance across a company. Tycom was pleased with the result and has, since, offered both Roman and Michal full-time positions.