Ugur Ali Kaplan

Machine Learning Master's Student
Machine Learning Researcher

In 2021, I have graduated as a computer engineer from the Istanbul Technical University and started my master's degree in Machine Learning at the University of Tübingen.

I worked in academic and industrial projects, in which I have taken part in all stages of ML workflow from data wrangling to model deployment.

My current work focuses on time series forecasting with uncertainty, using Gaussian Processes.

Experience

İTÜ Vision Lab Logo

Researcher

I am a member of İTÜ Vision Lab since October 2020 and work under the direct supervision of Prof. Dr. Gözde Ünal. Currently,

  • I work on time series forecasting on real-world data, in collaboration with domain experts at Eurasia Institute of Earth Sciences Ocean and Atmospheric Sciences Group at Istanbul Technical University.

In the past,

  • I used MLFlow to track experiments.
  • I used multivariate data imputation techniques to deal with missing features.
  • I used oversampling and undersampling techniques to work with unbalanced data.
  • I trained various models, such as logistic regression, SVMs, and gradient boosted decision trees.
  • I used Ray Tune and Ax for Bayesian Hyperparameter search.
  • I used Ray Serve to create a machine learning serving API.
  • I used SHAP to understand each feature's contribution to the model prediction.
Basira Lab Logo

Research Intern

Between April - October 2020, I worked as a research intern at Basira Lab. I worked under the supervision of Asst. Prof. Dr. Islem Rekik and graduate student Ahmed Nebli.

  • I optimized the code and significantly decreased the training time.
  • I contributed various ideas and implemented them.
  • I contributed to the paper writing process.
  • We published our paper at MICCAI PRIME Workshop and got the (workshop's) best paper award.
  • Model was made public at github.com/basiralab/EvoGraphNet/
inzva Logo

Tutoring

I worked as a guide for inzva at various times between September 2020 and Aug 2021.

  • I contributed to the development of the "Applied AI" program.
  • I developed lecture materials for "Applied AI" and "NLP Study Group" programs. Applied AI material can be accessed from inzva's GitHub repo.
  • I was selected the "Best Guide" in the NLP Study Group by the participants.