Asen Dotsinski
About me:
I am Asen Dotsinski, a PhD student at INSAIT since August 2026, where I am supervised by Dr. Yuxia Wang.
My research is on the security of open-weight large language models. I am driven by the goal of keeping increasingly capable AI safe as it proliferates, and I currently work on adversarial robustness against jailbreaking. I am also interested in tamper resistance and categorical unlearning, as complementary approaches to defence-in-depth.
If you would like to collaborate, feel free to get in touch.
Education
- University of Amsterdam, the Netherlands, September 2024 – July 2026
M.Sc. in Artificial Intelligence (GPA 8.65/10, Cum Laude, ELLIS Honors) - Eindhoven University of Technology, the Netherlands, September 2021 – July 2024
B.Eng. in Computer Science (GPA 9.2/10, Cum Laude)
Publications
- Asen Dotsinski, Panagiotis Eustratiadis
Sockpuppetting: Jailbreaking LLMs by Combining Prefilling with Optimization
Second Workshop on Agents in the Wild: Safety, Security, and Beyond (ICML 2026 Workshop) - Udit Thakur, Mohammad Hafeez Khan, Meher Changlani, Asen Dotsinski, Aswin Krishna Mahadevan, Ioannis Kechagias, Ivona Najdenkoska
CLaRE: CLIP with Latent Reconstruction Errors for Generated Face Detection
Proceedings of the 1st ACM Workshop on Deepfake, Deception, and Disinformation Security - Asen Dotsinski, Udit Thakur, Marko Ivanov, Mohammad Hafeez Khan, Maria Heuss
On the Generalizability of “Competition of Mechanisms: Tracing How Language Models Handle Facts and Counterfactuals”
Transactions on Machine Learning Research (TMLR); presented at EurIPS 2025
Relevant Experience
- Freelance Machine Learning Engineer – Motive Engineering, Remote (2024–2026)
- Back-end Developer – Drieam, Eindhoven, the Netherlands (2023 – 2024)
- Embedded Developer – WeFabricate, Best, the Netherlands (2022 – 2023)
- C++ Developer – Dimenco, Veldhoven, the Netherlands (2021 – 2022)
