Postdoc Researcher
Email: [email protected]
Dr. Dimitar Iliev Dimitrov
About me:
My name is Dimitar Iliev Dimitrov, and I am a Postdoctoral student in AI Security at INSAIT, advertised by Prof. Martin Vechev. My research interests span topics like Privacy and Robustness of Machine Learning Systems, Generative AI systems for low-resource languages, as well as improving AI reasoning systems. Previously, I obtained my PhD at ETH Zurich under the supervision of Prof. Martin Vechev on the topic of “Testing Federated Learning Privacy Through Gradient Leakage Attacks” and am part of the development team of BgGPT.
Education:
- ETH Zurich, September 2018 – May 2025
Direct doctorate student in Computer Science - The University of Edinburgh, UK, 2012 – 2016
BEng Computer Science - University of California, Irvine, USA, 2014 – 2015
Exchange student - Sofia High School of Mathematics, Sofia, Bulgaria, 2004 – 2012
Bulgaria diploma
2025
Maria Drencheva, Ivo Petrov, Maximilian Baader, Dimitar I. Dimitrov, Martin Vechev
GRAIN: Exact Graph Reconstruction from Gradients
In: International Conference on Learning Representations (ICLR 2025)
Kristian Minchev, Dimitar Iliev Dimitrov, Nikola Konstantinov
LARP: Learner-Agnostic Robust Data Prefiltering
In: International Conference on Machine Learning (ICML 2025) (Workshop)
Csaba Dékány, Stefan Balauca, Robin Staab, Dimitar I. Dimitrov, Martin Vechev
MixAT: Combining Continuous and Discrete Adversarial Training for LLMs
In: The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025)
Dimitar I. Dimitrov, Maximilian Baader, Mark Niklas Müller, Martin Vechev
SPEAR++: Scaling Gradient Inversion via Sparse Dictionary Learning
In: NeurIPS 2025 (Workshop)
2024
Kostadin Garov, Dimitar I. Dimitrov, Nikola Jovanović, Martin Vechev
Hiding in Plain Sight: Disguising Data Stealing Attacks in Federated Learning
In: International Conference on Learning Representations (ICLR 2024)
Dimitar I. Dimitrov, Maximilian Baader, Mark Niklas Müller, Martin Vechev
SPEAR: Exact Gradient Inversion of Batches in Federated Learning
In: Conference on Neural Information Processing Systems (NeurIPS 2024)
Ivo Petrov, Dimitar I. Dimitrov, Maximilian Baader, Mark Niklas Müller, Martin Vechev
DAGER: Exact Gradient Inversion for Large Language Models
In: Conference on Neural Information Processing Systems (NeurIPS 2024)
Timofey Fedoseev, Dimitar Iliev Dimitrov, Timon Gehr, Martin Vechev
Constraint-Based Synthetic Data Generation for LLM Mathematical Reasoning
In: Conference on Neural Information Processing Systems (NeurIPS 2024) (Workshop)
