I am Roel Leenders , a M.S. student in the Graduate School at PNU (Advisor: Prof. Myungho Lee). My current research interests are human-ai collaboration, machine learning.
Email : R.c.leenders@student.utwente.nl
CV | Google Scholar | Guithub | LinkedIn
Education
2023.03 - | Graduate school of Pusan national university(PNU)
2021.09 - 2023.06 | Graduate school of Pusan national university(PNU)
[Master] Elective coursework (Supervisor: Prof.), computer engineering
2016.09 - 2020.06 | Avans Hogeschool.
[Bachelor] Computer Science
2019.09 - 2019.12 | Minor Korean Studies, 부경대학교
[Exchange study] Korean language and culture
2015.09 - 2016.06 | Conservatoire of Music Amsterdam
[Bachelor] Guitar
Publications
[J1] Björn: An Interrogation Simulator
Roel Leenders, Pietro Camin, Ella Velner, Mariët Theune
In Proc. Frontiers in Artificial Intelligence and Applications , 2022
Work Experiences
2022.06 - | Software Engineer, Positive Digits
- Applying novel DevOps technologies like Terraform, Kubernetes and Docker to effeciently deploy and manage e-commerce platforms
2021.03 - 2022.06 | Product Owner, Fledge Solutions B.V
- Managed a team of 12 developers. Negotiated with stakeholders to realize their product requirements with the aim to update legacy ERP software to a fully cloud-based service.
- Taught 10 junior developers about Angular web development and Agile best practices
2020.02 - 2020.06 | Internship Software Engineer, Acorel B.V
- Helped to develop the product recommendation system of the largest hardware company in the Netherlands
2018.09 - 2018.02 | Internship Software Engineer, JouwWeb B.V
- Helped to develop the interface of a website-builder which is currently being used by more than 80.000 customers.
Research Experiences
Automatic data completion of breaking news events using Twitter :
- Building an event detection pipeline with the aim to automatically identify valid additional information about breaking news events.
Simulating interrogations using an embodied conversational agent :
- Modelled how the interpersonal stance of an agent would adjust based on the sentiment and question phrasing of the user’s utterances.
Predicting the progression of Pulmonary Fibrosis :
- Applied multi-task learning and 3D CNNs to improve the descriptive power of lung CTimages
Ensemble defense against FGSM and PGD attack on an instrument classifier :
- Improved the accuracy of the classifier by 20% after the ensemble defense.
Automatic story quality evaluation of generated novels :
- Extracted sentiment time-series from generated text and created a classifier to evaluate if the text contained a story arc.
Generation of clickbait headlines for data augmentation :
- Attempted to generate realistic clickbait headlines using both neural network and template based NLG methods.
Skills
Languages
English(Native)
Dutch(Native)
Korean(Fluent)
programming Languages
Python, JavaScript, Java, PyTorch
NLTK, Angluar, HTML, CSS, Docker, Koltin