Jesper Winsten
PhD Researcher | AI & Software Testing
Åbo Akademi University
Researching intelligent test automation with GANs. Passionate about machine learning, search-based software engineering, and powerlifting.
About Me
My research focuses on advancing software engineering through intelligent automation, specifically in the domain of cyber-physical system testing. I work on developing novel GAN-based methodologies to automatically generate test cases for complex systems, aiming to improve both efficiency and coverage in the testing process.
With a strong foundation in search-based software engineering and test automation, I'm driven by the challenge of solving complex problems at the cutting edge of software engineering research. My work combines theoretical research with practical applications, contributing to more robust and reliable cyber-physical systems.
Current Work
Åbo Akademi University
Developing novel GAN-based methodologies for automated test case generation in cyber-physical systems. Conducting research in search-based software engineering and test automation to improve efficiency and coverage in complex system testing.
Machine Learning Engineer
Developed machine learning solutions for welfare technologies, including an in‑house client welfare monitoring system based on pose detection. The system performs on‑device / edge inference and only transmits anonymized pose vectors and events (falls, unusual inactivity, posture changes), ensuring full privacy so no human can access raw camera feeds; providing real‑time alerts and aggregated, privacy‑preserving analytics for caregivers.
Education
Åbo Akademi University
Research focus on advancing software engineering through intelligent automation with GAN-based methodologies for cyber-physical system testing.
Åbo Akademi University & Reykjavik University
Double Degree in Intelligent Software Systems (CS), Nordic Master Programme in Intelligent Software Systems (NISS). The Nordic Master (NM) is the Nordic Council of Ministers’ funding programme,
which offers financing for joint Nordic Master’s programmes based on research, excellence and high quality. Master’s Programmes are executed by universities from at least two different Nordic countries
or their autonomous regions.
Thesis: Deep learning approaches for ship detection from inshore and offshore imagery.
Åbo Akademi University
Thesis: Utilization of connection sequences for connecting devices to fake Wi-Fi networks.