People

I research health record parametrization and optimization of clinical data collection processes that would help develop personalized decision-support systems for patients. I co-author data standardization framework for the Czech Republic.

I specialize in integrating, harmonizing, and FAIRification of real-world hospital data derived from biological samples stored in a biobank. As a data steward at the Masaryk Memorial Cancer Institute, I contribute to developing a metadata catalogue (data.bbmri.cz) which showcases available sequencing data linked to biological samples. I'm further interested in data exchange formats, ontologies, pseudonymization techniques, and developing efficient workflows for managing and sharing research data.

Adam has over a decade of experience in data science and brings knowledge from fundamental research in AGI as well as from the corporate. He led a research team at Seznam and improved its full-text and image search. From early-stage startups, he uses AI to disrupt industries such as real estate with automated price estimation and healthcare with precise measurement approaches—such as analyzing cough characteristics from real-world recordings using wearables. Currently at MindMed, he focuses on the use of LLMs to automatically evaluate HAM-A interviews with high precision. This enables consistent scoring and serves as a quality assurance layer for primary endpoints in clinical trials. Adam is also an educator (MLCollegue, Canonada), offering courses and mentoring companies. He is a founding member of the Brno Machine Learning Meetups.

I design human-centered AI for healthcare. My current goal is to empower patients with personalized tools that improve their health and well-being. I use HCD methodology, in which I have gained experience through freelancing, and I study information behavior of oncological patients for 5+ years. Before oncology, I was interested in decision support tools for criminalists and researched how to make network visualizations to support Criminal Police and Investigation Service of CR. Now I'm looking into decision-support systems for patients, and research LLMs for knowledge base organization, explainable recommendations, question-answering, and AI assistants. As a hobby, I create and teach generative art (generativedesign.cz, @zkrat).

I mainly research multimodal relational learning and its application for various biomedically relevant tasks such as drug repurposing or disease development prediction.

I research methods to improve computer readability of clinical notes. In my free time, I enjoy hiking, stargazing, and participating in puzzle hunts.