About sci2sci
We are a small team of experts who left their home domains to solve knowledge management in Life Sciences. We see this as the critical bottleneck — the thing stiffling progress in longevity, healthcare, quality of life and sustainability. We also believe that neither pure research nor pure product development alone can solve it. Both are necessary. That's why we founded sci2sci: a company that does research to optimize Life Sciences under the selective pressure of real-world constraints.
Our Founding Team
Dr. Angelina Lesnikova — CEO & Co-founder
Angelina is a neuroscientist who decoded neuroplasticity and memory at Eero Castrén's lab in Helsinki, then moved to Giuseppe Balistreri's viral cell biology group to research how SARS-CoV-2 enters the brain. She's lived the chaos firsthand: multi-site collaborations, sensitive data, evolving protocols, BSL facilities, and the constant question of "what was already done?" The problems sci2sci solves are the ones she experienced daily.
Valerii Kremnev — CTO & Co-founder
Valerii architected Fintech infrastructure — building trading platforms at Devexperts and engineering systems at Yandex and Delivery Hero that handle millions of transactions per hour under strict regulatory scrutiny. In trading and large-scale infrastructure, there are no compromises. The product either updates data multiple times per second on a mobile network in a developing country — or it's unusable. Traceability is not a nice optimization feature — it's a regulatory requirement. This background shapes how we build. Valerii brings the engineering discipline that makes safety architecture possible at scale.
It might sound like an unusual combination for a founding team — a neuroscientist and a fintech engineer. But we share two core insights:
- Simple mechanisms — in both biology and technology — can solve seemingly impossible problems when integrated coherently.
- Moving fast doesn't mean breaking things. It means acting with surgical precision.
How We Work
We have four core principles, balanced in pairs: Excellence and Pragmatism; Verifiability and Trust.
Excellence means we always meet the bar for necessary conditions. We don't sacrifice security or take shortcuts in mission-critical functionality. Our core backend is covered with 1000+ tests — regression, E2E and performance. We will be the first to push back on any exposure of our product's security surface. We don't allow direct write access to agents outside the sandbox. Mission-critical functions are done right.
Pragmatism is about sufficiency — and it allows us to survive excellence. With experience launching experimental functionality in Fintech and conducting research under pandemic pressure, we understand the reality of high-stakes environments. We don't just build for the ideal; we build for the inevitable. We know how to prioritize the ever growing backlog and find the optimal way to achieve project goals.
Verifiability is the core of our scientific and product expertise. Good hypotheses are falsifiable. We approach our customer journey as a hypothesis on the value of our product, and we structure our rollout phases to test it. If we can make any aspect of our product verifiable, we do.
Trust is equally important for our partnerships. We default to proofs over promises — a zero-trust mindset is in our DNA. But we also know that not everything can be verified upfront, and that real collaboration requires extending trust where proof isn't yet possible. So we earn it: through transparency, follow-through, and the same scientific rigor we apply to everything else.
What We Do
We build VectorCat — a data catalog and AI platform designed for biotech and pharma.
It connects the scattered reality of life sciences data: lab protocols in ELNs, sequencing files in cloud storage, spreadsheets on old network drives, BI dashboards, external CRO data. It creates lineage that maps both data processing and documentation — because in biopharma, you need full context.
We think FAIR principles (Findable, Accessible, Interoperable, Reusable) get the core insight right, but they're insufficient for broad adoption. They come from the academic world and don't survive the harsh constraints of real-world competition.
So we build systems that are SAFE: Secure, Auditable and designed to enable a FAIR Environment with appropriate constraints.
Search, governance, curation, AI-assisted research, data harmonization — all without migrating your data while keeping the strictest access policies and full auditability.
Our specialization is building simple, verifiable components that integrate into the ecosystem and amplify existing functions:
- Our audit trails are not logs — they are first-class evidence objects with full grounding and validation.
- Our AI agents are not separate functions — they are assistants to each user, inheriting their roles, permissions and context.
- Our search is not a vector store — it's aware of metadata and permissions, and distinguishes the right to find from the right to access.
These components interact with each other. Agents use search. Metadata carries audit evidence. Each component amplifies what others can do.
We believe products are not defined by a list of features for a marketing page but by the coherence of their components and their integrity toward solving the problem.