Computational R&D partner

Shortening the path from lab to life.

We help pharmaceutical and biotech teams turn complex biological data into decisions they can defend. Tailored algorithms, built on the biology, not around it.

Discover the expertises Our approach

Fig. 1 — How a model grows. An insect forewing, drawn vein by vein in the order the work happens.

01 · Wing base Biological data Omics, assays and clinical records enter the model at one source.

02 · Longitudinal veins Causal models The main paths from target to effect, carried as mechanism rather than correlation.

03 · Cell field Mechanisms tested Every cell between the veins is an assumption checked against the data, visible and accountable.

04 · Pterostigma The decision One weighted answer at the leading edge, with its uncertainty stated.

01 — What we do

Four things we are built for.

Simulate before you test

In silico environments and virtual patient models forecast drug behaviour and study outcomes before resources are committed.

Engineer trial intelligence

Adaptive, data-driven trial frameworks that adjust as the data arrives, improving efficiency and the probability of success.

Unify complex health data

Fragmented research and clinical sources become structured, analysis-ready systems that carry their own provenance.

Design systems as therapies

Algorithm-driven digital therapeutics and closed-loop systems where the software becomes part of the treatment.

02 — Why it matters

Nine in ten candidates never reach a patient.

Attrition concentrates in two places: a target that correlated with disease but never caused it, and a safety signal found after the money was spent. Both are failures of inference rather than chemistry, and both are reachable by mathematics.

Fig. 2 — Attrition network. Ten candidates enter at the source; nine branches end in a blunt terminal, one vein reaches a patient.

03 — Our approach

Four commitments behind every model.

01 Causal, not correlative

02 Mechanistic, not opaque

03 Uncertainty, always stated

04 Built for the regulator

Fig. 3 — Four commitments, four veins: each branches one more time than the last.

Read the approach

04 — Expertises

Eight needs, one engine.

Fig. 4 — One engine, eight terminals. Each branch is an expertise below.

Causal target discovery 02

Toxicity and safety prediction 03

Virtual patients and in silico trials 04

Delivery vector optimisation 05

Individual response prediction 06

Adaptive trials and real-world evidence 07

Surgery and drug-management efficiency 08

Closed-loop therapeutic systems

05 — Open Academy

Scientific tools should be understandable.

A growing library of free notebooks, simulations and walkthroughs for researchers and data scientists in the life sciences. Hands-on code in Jupyter, Python and R. No gatekeeping, no paywall.

Browse the library

Fig. 5 — The library, drawn where it exists and dotted where it is still growing.

Start with one decision you cannot make from the data you have.

We scope a first engagement around a single question — a target to validate, a trial to redesign, a safety signal to explain — and deliver a model you can interrogate, not a black box.

Fig. 6 — Many inputs, one answer.