Hi, I'm Soner

Data Scientist

I am a data scientist and software developer. I work on data analysis, machine learning, artificial intelligence, image processing and web development. For more information about me, please check the about section.

Details Let's Talk

About Me

my_photo

Data Scientist

Hello, I'm Soner Kar. I graduated from Eskişehir Osmangazi University with a degree in Electrical and Electronics Engineering and have been working as a software developer for 2 years. I use technologies such as Python, PostgreSQL, Node.js and React. In Python, I have worked in the fields of machine learning, deep learning, image processing, and data analysis. Last year, I started a Master's program in Artificial Intelligence at Ankara University and successfully completed the coursework, moving on to the thesis stage. On the web development side, I am working as a sole backend developer on a project that manages clients' GDPR processes using Node.js and PostgreSQL in an MVC structure. I have had the opportunity to work on large-scale projects, where I developed my skills in taking quick and accurate actions, producing useful solutions, and collaborating effectively with my team. My goal is to continue advancing in the software world by solving complex problems in artificial intelligence and related fields. You can contact me via my LinkedIn profile or by email.

View Details

My Journey

Education

2023 - ?

Master's Degree - Artificial Intelligence | English with Thesis

Ankara University

During my master's degree, I gained in-depth knowledge of artificial intelligence and machine learning. I have especially worked in the fields of deep learning and image processing and data analysis. Throughout my education, I focused on solving complex problems using large data sets and in this process, I worked with advanced tools such as Python, TensorFlow, Keras, Sklearn and libraries such as Numpy, Pandas, Sklearn Matplotlib, Seaborn, Plotly. Additionally, I improved my teamwork and problem-solving skills in my projects.

2018 - 2022

Bachelor's Degree - Electrical and Electronics Engineering | English

Eskişehir Osmangazi University

During my undergraduate education, I learned the basic principles of electrical and electronic engineering. I gained in-depth knowledge on topics such as circuit design, microcontrollers, signal processing and control systems. By taking part in various projects, I transformed my theoretical knowledge into practical applications. In particular, by working on artificial intelligence in my graduation project, I gained experience on EEG signals, deep learning and data analysis.

Experience

2023 - ?

Node.js Backend Developer

Felece

I worked as a Node.js backend developer at Felece Technology A.Ş. In this role, I developed applications that optimize various business processes and support digital transformation using SAP's cloud solutions. I gained proficiency in software development, MVC, RESTful APIs creation and integrations with Node.js. In this process, I used VsCode, PostgreSQL, Postman, Git, Github, Bitbucket and Jira tools. Additionally, I gained important experience in teamwork, problem solving and producing solutions for customer needs in projects.

2022 - ?

SAP C4C Developer

Felece

I worked as a developer on the SAP C4C (Cloud for Customer) platform at Felece. In this position, I developed applications that support and optimize customer relationship management (CRM) processes. I offered innovative solutions to customer needs by making customizations on SAP C4C. At the same time, I also gained in-depth knowledge of SAP C4C's integrations and data management. During this process, I improved my skills in working customer-oriented, paying attention to details and completing projects on time.

My Skills

Artificial Intelligence

  • Python

  • SQL, Postgre, Ms, NoSQL

  • Data Analysis

  • Machine Learning

  • Image Processing

    • Web Development

  • JavaScript

  • Node.js

  • React, HTML/CSS

  • SAP C4C & BTP

    • Tools

  • Jupyter Notebook, VS Code

  • pgAdmin, MS SQL, Postman

  • Kaggle, GitHub

  • Jira, Bitbucket

  • SAP Cloud App Studio, SAP BTP Cockpit

    • My Projects

    Click on the project you want to view.

    Mask Detector

    This project detects whether a face mask is worn or not. VGG16, a deep convolutional neural network model, was trained with the Transfer Learning method.

    (Click on the Live Demo button to use it)

    Live Demo


    Customer Segmentation

    Firstly, information was provided about Clustering, K-means Algorithm, Market Segmentation, and Customer Segmentation. Then, EDA (Exploratory Data Analysis) was performed on the Shopping Mall Customer Segmentation dataset, and customers were segmented based on their shopping habits and demographic characteristics using K-means algorithm.

    EEG Biometric Identification

    Routine EEG signals (brain signals) of 2.5 minutes were collected from 50 individuals. These signals were then decomposed into sub-bands, and an artificial intelligence classification model was built from scratch on the dataset created from these sub-band signals to determine which EEG signal belonged to which of the 50 individuals.


    House Sales Regression

    A regression model was developed to predict house prices using the house sales dataset from King County, USA. To evaluate the accuracy of the model, the R² score was calculated, and a high accuracy rate of 97.5% was achieved.

    Sumerian Cuneiform Image Processing

    In this project, the Sumerian Language Cuneiform dataset was used, and standardization and normalization were applied to all images. Then, image processing techniques such as blurring, division, thresholding (binarization), and morphology were applied to extract meanings from low-quality images.


    Glass Classification

    In this project, a machine learning model was developed to accurately classify types of glass using the glass classification dataset. Since the dataset is imbalanced, the performance of the model was evaluated with the F1 score, achieving an 80% success rate.

    Thompson Sampling on Ads

    Information was first provided about the Beta Distribution and Thompson Sampling algorithm. Then, the Ad Click-Through Rate Optimization dataset was used to determine which ads to show to users based on past click-through rates using Thompson Sampling.

    Contact Me!