Built from inside the reading room.
Verityn Ltd is registered in England and Wales. We are a London-based team of radiologists, machine learning engineers, and clinical quality specialists working on one problem: making the second read possible on every study.
It started with a finding that should not have waited.
Dr. Amara Nkomo spent thirteen years reading thoracic and emergency imaging in the NHS before founding Verityn. The pattern that prompted her was consistent: findings that surfaced on a formal second read had frequently been present, and visible, on an earlier film. Not missed through lack of skill. Missed because attention is finite and lists are long.
Tom Fairbrother had spent a decade building detection and localisation models and cared about one thing above correctness: calibration. A model that tells a clinician it is 94% confident should be right 94% of the time at that score. He had seen too many products that were not.
They started Verityn in March 2023 on a principle both agreed on: the tool has to show its reasons. A confident flag without a heatmap and a plain-language explanation is not useful to a working radiologist. It is noise.
Useful to the clinician. Honest about what it is.
Verityn is not trying to replace radiologists. It is trying to give every radiologist an instrument that reads alongside them without fatigue, without distraction, and without guessing.
Honest decision support
Verityn flags candidates and shows its reasons. It does not diagnose, and it does not report on its own. Any figure we quote is accompanied by the method that produced it.
The clinician decides
Every flag is a prompt to look again. The radiologist reviews, can dismiss, and signs the report. Verityn is in the loop, not in charge.
Calibrated and explainable
A confidence value should mean what it says. A heatmap should show what actually drew the flag. We hold ourselves to the standard of a clinical instrument, not a software product.
Privacy first
Identifiable imaging data can stay on your own infrastructure. On-premise and private-cloud deployment is available from the Practice plan, so data need not leave your environment.
Radiologists and engineers. Based in London.
14 people across clinical affairs, machine learning, regulatory, and engineering. All UK-based. Fictional for this preview, representative of the team Verityn is building.
Dr. Amara Nkomo
ClinicalChief Executive Officer, Co-founder
Consultant radiologist with thirteen years reading thoracic and emergency imaging in the NHS. Started Verityn after one too many findings surfaced on the second read.
Tom Fairbrother
Machine learningChief Technology Officer, Co-founder
Computer vision lead who has spent a decade on detection and localisation models. Cares about calibration, honest confidence, and models that show their reasons.
Dr. Sofia Marchetti
ClinicalHead of Clinical Affairs
Radiologist and clinical safety officer. Runs Verityn's reader studies and post-market surveillance.
Dr. Hannah Reyes
Machine learningPrincipal Machine Learning Engineer
Works on model calibration and heatmap faithfulness so a confidence value means what it says.
Priya Shah
RegulatoryHead of Quality and Regulatory
QARA lead steering Verityn's UKCA and CE technical files and its quality management system.
Daniel Okoye
EngineeringHead of Integrations
Builds the PACS-friendly connectors and the audit trail. Believes a good integration is one nobody notices.
Team profiles are illustrative. All individuals depicted are fictional and UK-based.
Verityn Ltd. Incorporated in England and Wales.
Headquartered at Third Floor, 24 Wharfdale Road, King's Cross, London. Funded by Beacon Health Ventures and two angels, a practising consultant radiologist and a former medical-device founder.
Founded
March 2023
Headquarters
London, United Kingdom
Team size
14 people
Market
Worldwide, UK launch
Registered
England and Wales
Quality system
ISO 13485 aligned
Four commitments we hold ourselves to.
We publish the method
When we quote a number, we describe how it was measured. Reader studies compare unaided reads against Verityn-assisted reads on de-identified retrospective datasets. The figures are in our validation summary.
We scope to what the model can do
Verityn reads the modalities it was trained and evaluated on. We do not generalise beyond the evidence and we are explicit about what is not yet in scope.
We keep the radiologist in control
The audit trail, the dismiss button, and the confidence heatmap all exist for the same reason: the radiologist must be able to interrogate, challenge, and override the model on every study.
We evolve with clinical feedback
Post-market surveillance is not a regulatory formality for us. What readers flag as unhelpful or wrong goes directly back into model and evaluation planning.
Give every read a second pair of eyes.
Verityn is decision support, built to the standard of a clinical instrument. The clinician always decides. Start free today, or talk to the clinical team.