About
Hi, my name is Milen. I am a postdoctoral researcher at the University of Copenhagen, Department of Computer Science, in the Software, Data, People & Society section. I work on the LEXplain project on explainability requirements for AI used in legal decision-making, in collaboration with the Faculties of Law at Copenhagen and Bergen. Here's something to keep you company as you read on: buna (ቡና)
Research
I research the explainability requirements for AI used in legal decision-making: what counts as an adequate explanation of a legal decision, how explanations differ from justifications, and what it takes for a decision to be contestable by the people it affects, namely citizens, case workers and courts. I work on these questions in the context of Nordic welfare administration, where automated decision-making is already in use and the legal requirements are concrete.
Before this, my PhD at the University of Amsterdam (CCI group) focused on normative control for data access in healthcare research: automating the GDPR and data sharing agreements using formal normative specifications in eFLINT, including purpose-based access control derived from the purpose limitation principle and duty lifecycle management for controller and processor obligations. I modeled the SIOPE DIPG/DMG data sharing agreement to enable fine-grained access control over rare disease clinical data.
Education
I received my PhD in Computer Science from the University of Amsterdam (April 2026) with the thesis Normative control for data access in healthcare research, supervised by Prof. Tom van Engers, Dr. Giovanni Sileno, Prof. L. Thomas van Binsbergen, and Dr. Dannis van Vuurden. I hold a double MSc in Computer Science from Saarland University (Germany) and University of Trento (Italy), where I studied Security and Privacy and minored in Innovation and Entrepreneurship through the EIT Digital program. I also studied technologies for active and healthy aging at the University of Trento for a year. I completed my Bachelor's degree in Computer Science at Jimma University (Ethiopia).
Teaching
I was a Junior Lecturer at the University of Amsterdam. I ran the lab sessions for the Human AI Interaction course, facilitating discussions and assignments, and supervised semester-long team projects in the Data Systems Project course. In the final year, I supervised five student teams, three of them on legal AI: linking legal advice letters to case law using retrieval-augmented generation, generating image descriptions for legal metadata with multimodal LLMs, and an AI agent for autonomous retrieval of legal case documents.
I also assisted in the following courses in the UvA Information Studies Master Program:
- Policy Making & Rule Governance
- Intelligent Interactive Systems
- Information Organisation
- Modelling System Dynamics
- Societal Complexity and Designing with Data
Thesis Supervision
I supervise master thesis students. Supervisions include:
- Assessing information quality in LLM-based pre-clinical health information systemsJaimy van Hattem
- Improving user comprehension of privacy notices through legal designEsref Paklaci
- LLM-generated narratives for comprehending EU AI Act regulatory conceptsRedouan Semlali
- Optimizing legal decision support for municipal governanceFloor Varkade
- Designing effective LLM-generated suspicious sign-in summaries: a requirements study with SOC personnelStijn Nagtzaam
Earlier supervisions (7)
- Representing access control systems using semantic webVan Beek, SF
- Analysis privacy policy languages and compliance to GDPRIlias Daia
- A generic consent management platform using eFLINTAbdul V. Zor
- An ontology based framework to support AI ActChristos Psaradellis
- Access control models in managing insider threatNiels Musters
- Security by design in software developmentSobbia Mohammad
- A business value perspective on digital twinsMaceo Linger