I’m an AI researcher specializing in Human-AI Interaction, AI in Education, and understanding the inner workings of Language Models. Based in Helsinki, Finland.
Currently a Research Fellow at Live Machinery, a non-profit AI safety research institute, where I work on human-AI interface design and collaborative sensemaking — developing frameworks for fluid interaction paradigms that reduce epistemic bureaucratization.
Research
- Comparing Behavioral Patterns of LLM and Human Tutors — ACL 2025, BEA Workshop
- Built an experimental framework comparing LLM teaching patterns (GPT-4o, Gemini Pro 1.5, LLaMA 3.1) with human tutors using the CIMA dataset. Identified systematic deviations in tutoring actions, response complexity, and behavioral patterns. Presented in Vienna.
- Concept Processing in LLMs — Ethical Landscaping, AI Safety Camp (2024)
- Architected a pipeline for systematic analysis of large language models. Developed methodology for statistical validation of token distribution patterns and extraction of complex relationships from model behavior data.
- AI in the Newsroom — Apart Research (2024)
- Analyzed 75,000 news articles to detect AI influence patterns. Identified statistically significant increases in ChatGPT-favored language. Published analysis.
- Probabilistic Modelling — University of Helsinki (2023)
- Enhanced Gaussian Process implementation combining neural networks with Stan. Research report.
Other Work
- DeepSeek V3 Analysis (2025) — Presented architecture breakdown at Tutke (Finnish Center for Safe AI)
- ML Bootcamp Creator (2022) — Designed and led an intensive 2-week ML program at Aalto University. Repository.
- Senior Mentor (2023–2024) — The Center for Effective Altruism. Guided AI safety student groups from TU Delft, Stanford, Groningen, Liverpool.
Education
Aalto University (2019–2024) — BSc & MSc in Artificial Intelligence, minor in Quantum Technology.