Daniel Trofimov
Independent biomedical AI researcher. Creator and maintainer of Robertium.
About
I build open-source tools for biomedical AI research.
My background is in software engineering, and I'm transitioning into biomedical informatics through hands-on project work. Robertium is my primary research project — a literature-based drug repurposing platform that applies Swanson ABC reasoning to surface cross-domain hypotheses from biomedical literature.
The catalog currently includes 1,618 outreach-quality candidates across ten therapeutic domains, with 856 novel candidates having no existing clinical trials. Full methodology is documented at /method, and the dataset is openly available under CC-BY-4.0 (DOI: 10.5281/zenodo.20110977).
I work independently, without institutional affiliation, with the goal of building methodology rigorous enough to be useful to researchers in established labs. If the work proves valuable, my long-term aim is to formalize collaboration with research groups in drug repurposing, oncology, and neurodegeneration.
Areas of focus
- Literature-based drug discovery (Swanson ABC reasoning, knowledge graphs)
- Cross-domain hypothesis generation in oncology and neurodegeneration
- Open data and methodology in biomedical informatics
- Retrospective benchmarking of computational drug repurposing methods
Open to collaboration
I'm actively interested in connecting with researchers working in:
- Pancreatic cancer biology, particularly KRAS-driven pathways
- Alzheimer's disease and EGFR signaling in neurodegeneration
- Drug repurposing methodology and validation frameworks
- Literature-based discovery and biomedical knowledge graphs
If any of Robertium's hypotheses align with your work, or if you have methodological feedback, I'd value the conversation. Reach out at daniel@robertium.com.
Links
- Robertium project — robertium.com
- Dataset on Zenodo — doi.org/10.5281/zenodo.20110977