OxiDiviner: A production-ready, open-source Rust library for time series analysis and forecasting, especially for financial markets. Features a wide array of models including ARIMA, GARCH, ETS, Kalman Filters, Markov Regime-Switching, and more. Offers multiple API layers for all expertise levels.

Jopara (Guarani-dominant mixed with Spanish) sentiment analysis corpus

Repository for systematic trial-and-error tuning of traditional machine learning models (Logistic Regression, SVM, Ensemble, etc.) without external libraries. Focused on improving validation accuracy through manual feature engineering, parameter adjustment, and custom implementations.

Geleneksel Makine Öğrenmesi Yöntemleri ile Çalışmalarım

Navigating technical interviews, I found many resources on basic algorithms either too complex or overly simplistic. As a machine learning engineer, blending algorithmic concepts with ML is key. This repo aims to balance complexity and practicality, making algorithmic challenges approachable and engaging.

2110446 Data Science and Data Engineering (2022/2) - Project

Handwritten Digit Classification on MNIST Dataset, Utilising Only Traditional Machine Learning Techniques and a Custom Feature Extractor, achieving highest accuracy of 98.08% with the same.

Repo for Meduzzen & U2D AI Internship challenges.

My code for practicals (or projects/labs) of course Machine Learning at Oxford.

Pushing things as I do data science stuff

A data science project analyzing NYC taxi trip durations using PySpark, from January to June 2019, including data preprocessing, geospatial visualization, and predictive modeling.

This project explores the usage of machine learning techniques in image denoising, particularly ridge regression and dictionary learning. It also includes an implementation of a readily runnable python script for capturing and denoising an image