Derisk Bio

Derisk Bio is a "bedside-to-benchside" company that uses a proprietary AI engine to analyse real-world patient data first, creating a de-risked and predictable path for developing new biomarkers and therapies.

Problem

The Discovery Problem: A Failure to Learn from Biology

It costs $2.6 billion to bring a single new drug to market, yet 97% fail in human trials. This is the result of a discovery model fundamentally disconnected from biological reality. The traditional "benchside-to-bedside" approach is a linear process, testing simplified lab theories against the complexity of human disease, with no systematic way to learn from failures or predict success.

The core issue is a data problem. Human biology generates massive, multi-layered datasets—from clinical outcomes and genomics to cross-species experiments—that contain the clues to why a therapy works for one patient but not another. The traditional model lacks the tools to integrate and interpret this complexity, forcing researchers to rely on siloed information and incomplete hypotheses. The critical signatures that could guide development remain buried, and predictable science remains out of reach.

To solve our hardest diseases, we must move beyond this one-way street. We need an engineered, iterative process that can navigate biological complexity and learn directly from patient data. We need a new discovery engine built for the modern data landscape, capable of transforming a high-risk endeavour into a predictable path to the clinic.

Solution

At Derisk Bio, we solve the discovery problem by flipping the traditional model. We don't start with a simplified lab theory; we begin with the biological reality of the patient. Our "bedside-to-benchside" philosophy is built on a powerful, iterative loop: we analyse patient data to generate novel insights, use those insights to guide focused pre-clinical experiments and then validate our findings back in patient cohorts. This cycle is repeated until we achieve a robust, translational outcome that is ready for the clinic.

Powering this process is the Recursiv Engine, our proprietary AI discovery platform. We have already built our foundational AI methods and our first foundational model, which deploys different specialised algorithms at each step of the discovery cycle. The engine integrates and learns from massive, complex datasets—from large-scale clinical cohorts to cross-species experimental data—to navigate the complexities of human biology and pinpoint the signatures that matter.

This unique combination of an iterative discovery process and purpose-built AI transforms high-risk research into a predictable science. By engineering the path from patient to lab and back again, we de-risk development, identify the right patients for the right therapies and uncover the novel biology needed to create the next generation of cures.

Contact

Jit Sarkar

James Arnold LinkedIn