Prague-based biotech research firm Theorema, which develops an AI platform to automate scientific R&D workflows, has secured €2.85 million (CZK 70M) in pre-seed funding to expand its autonomous research platform.
- Established in 2025 by Filip Doušek and Hynek Walner, Theorema develops an AI platform to automate scientific research and R&D workflows. The system converts research questions and raw data into multi-agent processes that review literature, analyze datasets, and design experiments across biotechnology, pharmaceuticals, chemical manufacturing, and advanced materials.
- The platform has two main functions. The Data Analyst agent handles routine analysis, generating tables, visualizations, and interpretations while working with existing data systems. The R&D Forensics module examines terminated programs and negative datasets to identify salvageable directions, prevent redundant experiments, and support research planning.
- Theorema agents cover the entire research cycle, including hypothesis mapping, in silico modeling, experiment design, laboratory execution, and decision support. Each run updates a cumulative model of mechanisms, hypotheses, and results. The system is applied in genomics, proteomics, enzyme engineering, reaction design, and materials science, focusing on complex datasets and high-cost experiments.
- Data ownership remains with the user, and deployment can occur in a secure cloud or private infrastructure. All analyses are versioned and reproducible, providing a full audit trail for regulatory and intellectual property purposes.
Details of the deal
- Theorema’s €2.85 million pre-seed round was supported by Tilia Impact Ventures, InvestEU, Credo Ventures, KAYA VC, Amino Collective, NAVEC Investment Management, and i&i Biotech Fund.
- The firm plans to accelerate the development of its AI-driven platform, enabling autonomous management of the entire scientific R&D process, from literature search and experiment design to real-time lab control and data analysis, ultimately compressing research timelines from weeks or months into hours.
