sci2sci secures €1.2M to develop trusted AI infrastructure for regulated industries
Berlin-based software startup sci2sci has raised €1.2 million in pre-seed funding to expand its technology for organising, connecting and verifying data in regulated industries. The round was co-led by Heliad and IBB Ventures, with participation from Robin Capital and Superangels.
Founded by Angelina Lesnikova and Valerii Kremnev, sci2sci is building software that helps companies connect fragmented information and verify the evidence supporting data and AI-generated results.
The company initially focused on biopharma, where critical research and regulatory information is often scattered across PDFs, spreadsheets, laboratory notes and disconnected systems. This fragmentation can make it challenging to trace findings back to their original sources, especially during regulatory reviews and audits. AI tools can further increase these challenges when they rely on incomplete or inconsistent information.
Sci2sci's Integrity Cortex is designed to address this issue by connecting company documents, data and AI-generated outputs into a structured network. The platform links claims to their underlying sources and verifies citations and conclusions, creating an auditable record intended to meet 21 CFR Part 11 requirements for electronic records. Integrity Cortex is built on Parseltongue, a framework that sci2sci open-sourced this year under the Apache 2.0 licence.
The startup's second product, VectorCat, acts as a data integration layer that connects information across cloud storage, network drives and laboratory systems without requiring companies to move their existing data. It creates a searchable catalogue that can be used by both employees and AI applications.
Valerii Kremnev, co-founder and CTO of sci2sci, explained that the company's approach is designed to build verification and traceability into AI workflows:
Instead of shipping capabilities first and patching trust and security afterward, we built Parseltongue and Integrity Cortex to permit only safe outputs. A model’s lockpicking skills are useless when there’s no door — and our software guarantees there is none. The same holds for deception: our systems don’t have ungrounded statements as a primitive. The model is forced to produce evidence for every claim it makes.
Sci2sci is already working with customers across preclinical research, contract clinical research and bioprocess operations. The new funding will support the expansion of its engineering team, strengthen existing customer deployments and help the company reach more biopharma customers. Sci2sci also plans to expand Integrity Cortex into other regulated sectors, including banking.