Core
A common foundation for physical models, data and evidence. Assumptions, provenance and validity limits stay connected to each result.
calyr / CODE-BASED MODELS · SCIENTIFIC AI
calyr.aí
Surrogate models.
Better decisions.
calyr connects physical models, fast surrogate predictions and experimental evidence to guide the next decision. The platform is in development, with uncertainty and validation made explicit.
Explore the platform01 / THE PLATFORM
One foundation.
Many scientific questions.
calyr Core, Engine and Model Capsules form the common architecture. Applications build on this foundation.
The aim is to make complex scientific models useful in everyday research: faster comparisons, explicit uncertainty and a traceable reason for the next experiment.
Engine
Runs models and coordinates prediction, comparison and optimisation. Surrogates make expensive calculations faster; oracles help select the next informative test.
Model Capsules
Reusable, domain-specific model packages with defined inputs, outputs and evidence requirements. Each capsule must be evaluated for its intended use.
Development status: a research and software platform under construction. Predictive performance and validation must be established separately for each model and use case.
02 / THE METHOD
Predict. Test.
Learn from evidence.
A prediction supports a decision only when its assumptions, uncertainty and evidence are visible.
Define the question
Specify the objective, available evidence and constraints. Make clear what a useful answer would change.
Predict and compare
Use physical models and learned surrogates to explore alternatives. Report uncertainty and where the model applies.
Test what matters
Select a simulation or experiment that can resolve the important uncertainty. Use independent evidence to check the prediction.
03 / APPLICATIONS
The platform,
put to work.
Commercial tracks and research domains explore different uses of the same modelling foundation.
01aorta / aflowtexPatient-specific flow modelling and intervention design. A research and validation domain for the platform.02bríxGeometry, simulation and surrogate models for engineering design exploration.03biophysical inferenceSPR, SAXS and nanoparticle models connect measurements to competing physical explanations.
Further exploration: trajectoríes
An exploratory calyr track: compute individual particle trajectories and optimise purification protocols. The planned implementation uses JAX for differentiable computation. LNP / mRNA is the first use case; experimental validation remains ahead.
calyr develops scientific models, surrogates and inference. LITHÍOS applies this foundation to physical products and design questions.
04 / RESEARCH & EVIDENCE
Follow the work.
Methods, research directions and decisions with their current evidence status.