About Me

I am currently a first-year Computer Science MSc student @ ETH Zürich, majoring in Machine Intelligence with a minor in Computer Graphics. I graduated from Sabanci University in 2024 ranked $\text{1}^{\text{st}}$ in both my program and my faculty. I have previously contributed to a project about the evasion attacks on CNN models, together with Xiayoun Xu from Radboud University Nijmegen as an Erasmus+ Intern. I am almost done with my graduation project, which is about habit formation in gym customers, using various techniques from network science, machine learning, and data analysis. We hope to publish this project soon in one of the esteemed journals.

For collaboration opportunities or any other inquiries, please contact me via email.

Research Interests

• Machine Learning
• Evasion Attacks on Deep Learning Models
• Network Science

Selected Projects

• From Occasional to Steady: Habit Formation Insights From a Comprehensive Fitness Study
  
  Efe Tüzün, Ege Demirci, Ahmet Furkan Ün, Taner Giray Sönmez
  
  This study utilizes data from over 250,000 customers and approximately 33 million visits collected by Mars Athletic Club, during the period from 2022 to 2023. Our research provides a comprehensive analysis of the critical factors that influence the formation of consistent exercise habits among gym members. Through our survival analysis, we identified key milestones in habit formation where the likelihood of maintaining regular attendance declines significantly.
  
  Paper (soon to be available)

• Decoding Istanbul’s Vehicular Veins - A Network Science Exploration of Metropolitan Traffic Dynamics
  
  Efe Tüzün, Ege Demirci, Özgür Yılmaz Beker, Alp Tüzün
  
  Using network science techniques, we analyzed Istanbul’s traffic patterns. We identified key congestion points, proposed a spatial correlation model, and offered insights for urban transport planning in major cities.
  
  Paper | GitHub

• Political Tweet and Bot Classifier - Study on Turkish Twitter
  
  Efe Tüzün, Ege Demirci, Alp Tüzün, Deniz Can Gezgin
  
  The project aims to address two issues in the context of Turkish Twitter: the proliferation of political content and the prevalence of bot accounts. Utilizing advanced machine learning techniques, the project seeks to classify political tweets and distinguish between genuine user accounts and bots.
  
  GitHub

Teaching Experience

• Learning Assistant for CS 412 - Machine Learning @ Sabanci University (Oct 2023 - Jan 2024)
• Learning Assistant for IF 100 - Computational Approaches to Problem Solving @ Sabanci University (Oct 2021 - Jun 2022)