Taner Giray Sönmez

I am a Master of Science candidate at Columbia University, now in my final semester specializing in Machine Learning. This 2025 summer, I interned as a Software Development Engineer at Amazon Web Services (AWS) on the AWS Glue team. There, I enhanced Personally Identifiable Information (PII) detection by deploying a new ML model that reduced false negatives by 37% and false positives by 20%.

My coursework at Columbia includes Natural Language Processing, Artificial Intelligence, Neural Networks and Deep Learning, and High-Performance Machine Learning. I have also contributed as an AI Researcher at the Columbia University IRT Lab, where I engineered an LLM-based system for privacy policy analysis, achieving 99% weighted recall in clause classification using a Retrieval-Augmented Generation (RAG) pipeline.

I graduated in June 2024 from Sabanci University with a B.Sc. in Computer Science and Engineering. During my undergraduate studies, I completed an Erasmus exchange program at Uppsala University in Sweden. My final year project at Sabanci focused on exercise habit formation, using machine learning and causal inference to analyze data from Turkey’s largest gym chain. You can access our paper via this arXiv link.

Throughout my studies, I’ve also gained industry experience as a part-time Data Scientist at Mars Athletic Club and academic experience as a Learning Assistant for the Computational Approaches to Problem Solving course at Sabanci University.

Interests

• Machine Learning
• Natural Language Processing
• High Performance Computing