The best way to verify that you are alive is by checking if you like variations.
- Nassim Nicholas Taleb
Hi, I’m Magret.
I’m a Software Engineering graduate with a focus on machine learning and mathematical modeling.
I believe that a solid mathematical foundation is key to building meaningful, interpretable, and reliable models. This principle shapes both my research and the way I design software.
As a Teaching Assistant, I enjoyed leading hands-on tutorials and always aimed to make abstract concepts intuitive and applicable.
Outside of work, I love baking cakes, reading non-fiction, and staying active through running and cycling.
Feel free to explore my portfolio to learn more about my projects, research interests, and PhD aspirations.
Research
Masters' Thesis
Olive trees hold significant cultural and agricultural value in many Mediterranean regions, including Cyprus, where I completed my master’s. My thesis focused on building a dynamic forecasting model for temperature and growing degree days (GDD), tailored to the needs of olive cultivation.
At the core of the model is a Bayesian Ensemble Kalman Filter (BEKF), which updates an ensemble of GRU-based neural networks in real time as new weather data arrives. This approach enables uncertainty-aware predictions and allows the model to adapt to shifting climate conditions. The filter improves robustness by capturing both model and observational uncertainty, making it more reliable for decision-making in agriculture.
Link to github: https://github.com/Magret-Oladunjoye/uncertainty-aware-forecasting
DeepFarm
DeepFarm is an EU-funded project that brings AI and big data into agronomy education. It connects universities in Europe, Madagascar, Côte d’Ivoire, Haiti, and the Dominican Republic to close the digital skills gap in agriculture. I work with Prof. Dr. Hadi Işık Aybay to design and teach lab sessions focused on helping students analyze agricultural data and apply machine learning.
As part of this work, I co-authored a conference paper for EDULEARN25 and am currently collaborating on a research paper focused on YOLOv8-based severity detection of the olive peacock spot disease. Our approach combines leaf-level classification with early-lesion-aware attention mechanisms.
Link: deepfarm.eu
Some Projects:
-
Analyzing rainfall variability and climate patterns in Ondo State, Nigeria, using stochastic methods
Teaching
During my master's at Eastern Mediterranean University, I was a Full-time Teaching Assistant. Below are some of the courses I supported:
-
Foundations of Computer-Software Engineering: An Introduction to Python (CMPE107)
-
Object Oriented Programming (CMPE211)
-
Operating Systems (CMPE242)
-
Principles of Programming Languages (CMSE318/CMPE410)
Volunteering
I dedicate some of my spare time to volunteering at my local animal shelter, where I engage with the dogs and lead initiatives to enhance their welfare.
References
For inquiries related to research, collaborations, or professional opportunities, feel free to reach out.
You can either fill the contact form or send me a direct mail at: magretoladunjoye@web.de
