Антон

Anton Kisilishyn
Kaggle Competitions Expert kaggle.com/antonius42

Ukraine, Cherkasy | [открыть контакты](см. выше в блоке «контактная информация») | [открыть контакты](см. выше в блоке «контактная информация») | github.com/antk42

Education

Fachabiture in Felix-Klein-Gymnasium Göttingen, Deutschland 2015-2017
Data Science student in ChSTU, Cherkasy, Ukraine 2022-present

Machine learning and data analysis: Online courses from MIPT and Yandex on Coursera
Karpov Courses
Open Machine Learning Course

Languages: English, German, Ukrainian, Russian

Skillls:

Programming: Data visualization: Modeling:
Python(Pandas, Numpy, Matplotlib, Seaborn, Logistic regression, Linear
Scikit-learn, XGBoost, Power BI regression, Decision trees,
CatBoost, LightGBM, Random Forest,
Tenserflow, Torch, Gradient boosting,
TorchVision, TorchText, Neural Network(CNN,FNN,
NLTK), SQL, Java, Golang, RNN)
Linux ,Git

Experience:
Independent Freelancer 2020 - current

Leveraged advanced web scraping techniques using BeautifulSoup, Scrapy (Python), and Golang
to extract and analyze intricate datasets from diverse websites.

Kaggle Competitions:

Enefit - Predict Energy Behavior of Prosumers
• Achieved: Silver Medal (Ranked 78th out of 2731 participants)
• Developed a model to predict energy behavior of prosumers, addressing the challenge of energy
imbalance and its associated costs. Prosumers, who consume and generate energy, significantly
contribute to this imbalance, posing logistical and financial challenges for energy companies.
The Learning Agency Lab - PII Data Detection
• Achieved: Bronze Medal (Ranked 174th privately, 52nd publicly out of 2048 participants)
• Partnered with Vanderbilt University and The Learning Agency Lab to develop automated
techniques for detecting Personally Identifiable Information (PII) in educational datasets.
This initiative aims to enhance data privacy by screening and cleansing educational data
before public release, enabling the creation of high-quality open datasets for educational
research and intervention.
HMS - Harmful Brain Activity Classification
• Private Rank: 371st out of 2767 participants. Public Rank: 287th out of 2767 participants.
• Developed a model to detect and classify seizures and harmful brain activity using
electroencephalography (EEG) signals. This work holds potential for significantly improving
neurocritical care, epilepsy treatment, and drug development by enabling faster and more accurate
detection of brain abnormalities.