By way of
the machine.
A UK research and engineering company. Computational science, machine intelligence, and the engineering that carries both into service.
Four divisions. One method:
measure it, then build it.
Computational Science
Molecular simulation and quantum chemistry, orchestrated by AI.
Quantum chemistry, molecular dynamics, force field parameterisation and synthesis planning, assembled into automated pipelines that rank candidates before anyone runs an experiment.
Machine Intelligence
Learning algorithms that continue to learn after training stops.
Fundamental research into how learning systems assign credit, retain what they already know, and what any of it costs in energy. Run as a research programme, with hypotheses fixed before the work starts.
Applied Engineering
The distance between a result and a system people can use.
Production platform build, data and compute infrastructure, HPC and GPU pipelines, and the integration work that turns a validated method into something a team runs on a Tuesday.
Defence & Security
Sovereign delivery into environments that cannot phone home.
Containerised, air-gappable deployment with zero external dependencies. Threat modelling, accreditation support, and applied work across CBRN, biodefence, energetics and autonomous coordination.
Molecular Simulation
Quantum chemistry, molecular dynamics and force field parameterisation, orchestrated by AI into pipelines that screen and rank candidates before the first experiment is run.
Learning That Continues
Credit assignment without a backward pass through the whole network. Learning rules built from information already present where the change has to be made. That is what lets a system keep learning after deployment, on a fraction of the power.
Sovereign Delivery
Built for networks that cannot phone home. Containerised, air-gappable, zero external dependencies. UK sovereign, Five Eyes aligned, accreditable by design.
Prius mensura, deinde machina.
First the measure, then the machine
Dr. Adam Sykes
Founder & Chief Scientist
Two decades across computational chemistry, software engineering and AI systems. Via Machina puts the science and the production engineering under one roof, divided so that each kind of problem goes to the discipline that can actually answer it.
A small permanent team holds the research agenda and the platform code. Continuity sits with them, so nobody has to brief a fresh bench every quarter.
Divisions draw on named specialists for each programme: chemists, ML researchers, engineers, security practitioners. Each one is engaged for the work they are the right person for.
Every division works the same way: state the claim, fix the test before running it, record the result whichever way it falls, then build only what survived.
Get in touch.
Tell us which division your problem belongs to. If you are not sure, describe the problem and we will tell you. One address reaches all four.
hello@viamachina.ai