Hi, my name is Abdullah KAVAKLI
I'm a Data Scientist.

Know more

About me

Profile Image

I am a data scientist and have knowledge in DevOps and machine learning. Skilled in implementing Linux projects on the cloud, feature engineering, and various machine learning algorithms. Developed successful machine learning models with high accuracy scores. Experienced in NLP techniques for sentiment analysis. Optimized algorithms for faster processing and am passionate about using data science to solve complex problems and help businesses make informed decisions. I'm currently pursuing a master's degree in computer engineering at Istanbul Technical University, a prestigious university in Turkey.

View Resume


Credit Card Fraud Detection using Machine Learning

  • Developed a machine learning model to detect fraudulent credit card transactions using the Python programming language and the Scikit-learn library.
  • Achieved a ROC-AUC score of 98% in detecting fraudulent transactions, outperforming traditional rule-based fraud detection systems.
  • Cleaned and preprocessed a real-world dataset containing credit card transactions to identify patterns in the data.
  • Utilized various feature engineering techniques such as scaling, oversampling, undersampling, and power transform to optimize the model’s accuracy and performance.
  • Trained and fine-tuned multiple machine learning algorithms to identify the best model for the dataset.

Source Code

Brain Tumor Classification with CNN

  • Developed a deep learning model using Convolutional Neural Networks (CNNs) to classify brain tumors from MRI images.
  • Implemented the model using Python and TensorFlow, achieving a ROC-AUC score of 95% on the test set.
  • Used transfer learning by fine-tuning a pre-trained InceptionResNetV2 model, which significantly reduced the training time and improved the accuracy of the model.

Source Code See Live

Restaurant Reviews Classification

  • Utilized NLP techniques to preprocess and clean a dataset of restaurant reviews.
  • Engineered features from the text data to represent sentiment and subjectivity.
  • Utilized the Scikit-learn library to train and test ML algorithms.
  • Achieved a high ROC AUC score of 93% on the test set.

Source Code

Stock Market Prediction Using RNN

  • Developed a model for predicting stock prices by training LSTM and GRU models on historical data from the Yahoo Finance API.
  • Utilized the TensorFlow library for model development and performance evaluation.
  • Visualized model predictions and achieved a Root Mean Square Error (RMSE) of 2.86 on the test set.

Source Code

Accelerating the PrefixSpan algorithm using GPU

  • Developed and implemented a GPU-accelerated version of the PrefixSpan algorithm, a sequential pattern mining algorithm.
  • Achieved a 75% performance improvement by using the Numba and Numpy libraries.


Feel Free To Contact Me Anytime

Get in Touch