CytoLearn [DEFUNCT]

CytoLearn was a startup I co-founded and served as CTO, developing AI-powered flow cytometry analysis software for blood cancer diagnosis. The platform used machine learning to learn from expert operators’ manual gating decisions, enabling automated analysis that was 30x faster than the gold standard method while maintaining 99.5% accuracy after training on just 10 samples.

The software was designed to address a critical bottleneck in NHS diagnostic pipelines, where manual flow cytometry analysis can take 15 minutes per sample and requires highly skilled operators. CytoLearn captured lab-specific knowledge so it wouldn’t be lost when staff left, while allowing the rest of the diagnostic pipeline to remain unchanged.

Despite securing investment and making significant technical progress, the company ultimately shut down after a few years due to intellectual property law complications with the University of Liverpool. Despite extensive negotiations, we were unable to reach an agreement that would allow the project to continue.