Middle East Technical University
Sep 2026 – PresentM.S. in Computer Engineering · Ankara, Türkiye
Effect-Driven Skill Abstractions for Offline Reinforcement Learning
NeurIPS 2026Fortieth Annual Conference on Neural Information Processing Systems
Burcu Kılıç
I am a master's student in Computer Engineering at Middle East Technical University, where I work with Assoc. Prof. Emre Ugur at ROMER. My research interests lie at the intersection of robot learning, reinforcement learning, and learned abstractions for planning.
I received my B.Sc. in Computer Engineering from Bogazici University, where I was a research assistant in the Cognitive Learning and Robotics (CoLoRs) Lab. I also studied at KTH Royal Institute of Technology as an exchange student and worked as a visiting researcher at the University of Innsbruck with Justus Piater.
I am interested in how robots can acquire useful skills and compact representations from interaction data, then use those representations for decision-making and planning.
Publications
See also Google Scholar.
Joint Discovery of Object and Action Symbols through Effect Prediction for Robotic Manipulation Planning
arXiv 2026Preprint · 2026
Trajectory Learning with Graph Representations for Social Robot Navigation
arXiv 2026Preprint · 2026
Education
Bogazici University
Oct 2021 – Jan 2026B.Sc. in Computer Engineering · Istanbul, Türkiye
KTH Royal Institute of Technology
Jan 2024 – May 2024Exchange student in Computer Science and Engineering · Stockholm, Sweden
Experience
Selected highlights. See my CV for more details.
Middle East Technical University
Sep 2026 – PresentGraduate Research Assistant · Ankara, Türkiye
- Conducting research with Assoc. Prof. Emre Ugur at ROMER.
PragmaCraft
Apr 2026 – Jul 2026Data Scientist · Istanbul, Türkiye
- Developed document search and retrieval-augmented generation features for Answie.
Cognitive Learning and Robotics (CoLoRs) Lab, Bogazici University
Dec 2024 – Jan 2026Research Assistant · Istanbul, Türkiye
- Researched robotic skill discovery, learned symbolic abstractions, and planning from interaction data.
- Trained deep learning models using simulated and real-robot data.
MYTH AI
Aug 2025 – Sep 2025Computer Vision Intern · Istanbul, Türkiye
- Built garment-compositing pipelines using depth estimation, segmentation, and OpenCV.
University of Innsbruck
Jun 2025 – Aug 2025Visiting Researcher · Innsbruck, Austria
- Implemented Decision Transformer in PyTorch and studied RL methods on D4RL and OGBench.
MTU Aero Engines
Jun 2024 – Sep 2024Machine Learning Intern · Munich, Germany
- Developed LLM-based agentic systems with LangChain in Azure ML.
- Built PostgreSQL-backed dashboards and applied data analysis and forecasting.
ModaMind
Jul 2023 – Jan 2024AI Developer · Istanbul, Türkiye
- Fine-tuned Stable Diffusion with LoRA and ControlNet for virtual-fashion applications.
- Developed retrieval-augmented generation pipelines with Llama and GPT models.