Language

Batu KorayMasak(Last updated · August 2026)

Undergraduate researcher at Koç University’s Security, Privacy and Data Engineering (SPADE) Lab and double-major student in AI & Data Engineering and Computer Science at Özyeğin University.I currently conduct research on Membership Inference Attacks in the domain of Large Language Models at SPADE Lab, a topic in AI Security that is becoming more important by the day.

Current Track

What I'm working on right now.

  1. Undergraduate Researcher at Koç University

    Working in the SPADE (Security, Privacy and Data Engineering) Lab at Koç University under Asst. Prof. M. Emre Gürsoy, researching membership inference attacks against fine-tuned autoregressive language models. The question is basically: can you tell whether a given piece of text was in a model’s training data just from how it behaves? Turns out they’re worse at keeping secrets than they’d like you to think.

    I build fine-tuning datasets, fine-tune models on HPC clusters, run the attacks and report metrics like AUROC and TPR@1%FPR, and help develop NLP and statistics-driven attack methods. I present weekly results to the lab, and I'm very grateful for Prof. Gürsoy's mentorship.

  2. July 2026 - Present

    Teaching Assistant at Özyeğin University

    I give face-to-face and remote peer-teaching sessions through Özyeğin University's Office for Learning and Teaching Enhancement (OLTE), with problem-solving support and exam preparation for mathematics and engineering courses. In Summer 2026 I taught MATH103 (Single-Variable Calculus), MATH104 (Multi-Variable Calculus & Infinite Series), MATH212 (Differential Equations) and MATH217 (Probability & Statistics), and in the 2026–2027 Fall term I'm teaching Probability & Statistics and Data Science.

    I do this for two reasons: because I love mathematics, and because teaching the fundamentals forces me to understand them more deeply.

Academic Statistics

A quick view of academic standing and the work behind it.

3.77

CGPA / 4.00

Maintained while double-majoring in AI & Data Engineering and Computer Science.

3 / 65

AI & Data Engineering

Consistently ranked in the top 3 of my department throughout university.

3%

Engineering Faculty

Top 3% across 1,727 engineering students.

4 / 4

High Honor semesters

Performance sustained across every completed semester.

Education

A double-major foundation in AI, data engineering, and computer science.

2024 – Present

B.Sc. AI & Data Engineering · B.Sc. Computer Science

  • CGPA3.77 / 4.00
  • HonorsHigh Honor status across all four completed semesters
  • Rank3rd of 65 in department · top 3% of Engineering Faculty
  • RolesTeaching Assistant · Peer Advisor (2025–2026)
Issued by Özyeğin University · June 2026

High Honor Standing

Achieved High Honor academic standing in all four semesters to date, clearing the required ≥3.50 GPA bar with zero failing grades every single term.

Issued by Özyeğin University · Aug 2025 and Aug 2026

2× Academic Merit Scholarship

One of three awarded across ~300 eligible students in the Faculty of Engineering for 2025–2026, ranked purely on cumulative GPA. Received the scholarship again for 2026–2027 after meeting the academic merit conditions for 2025–2026. Each award provides a 25% tuition waiver.

Özyeğin University · Sep 2025 to June 2026

Peer Advisor

Mentored approximately 20 first-year AI & Data Engineering students, maintained a curated GitHub repository of course materials and academic guides, and ran an active advising channel for coursework, programs, and research opportunities.

ETS · February 2026

TOEFL iBT · 110 / 120

CEFR C1 proficiency, with perfect section scores in Listening and Writing.

2019 – 2024

High School Diploma

2020 to 2023

FIRST Robotics Competition, Team 8151 Wildfire

Spent three years on my high school's FRC Team 8151, Wildfire, from 2020 to 2023. I joined the Programming subteam in my first year and spent the following two years on the Electronics subteam. Alongside the technical work, I was usually the one keeping the team laughing when things got stressful.

Team 8151 Wildfire on Instagram ↗

Programming coursework

My first formal programming courses were in high school, where I learned Java and Python.

Research Focus

Four directions shaping my research.

1

Privacy and membership inference in language models

Membership inference attacks on fine-tuned autoregressive language models, especially methods that determine whether a given text sample was part of a model's training data from its output probabilities and loss behavior.

2

Adversarial attacks and LLM red-teaming

Multi-turn jailbreaking and adversarial prompting strategies that exploit persuasion, context accumulation, and conversational framing, with an emphasis on understanding why safety fine-tuning fails under sustained interaction.

3

NLP and transformer architectures

Language modeling, attention, transformer internals, and the mechanisms underlying the systems I study from both the implementation and adversarial sides.

4

Language models in psychology

How much psychologically meaningful structure can language models recover from text, from mood and affective signals to personality traits and clinically relevant linguistic patterns.