Noetik is an AI-native biotechnology company based in South San Francisco, California, working on precision cancer treatment. The company trains foundation models of human tumours and tissue microenvironments by combining multimodal tumour data with self-supervised learning. Its stated aim is to apply machine learning to practical oncology problems - target discovery, biomarker development and prediction of therapeutic response - rather than to generic benchmarks.
Its technology stack spans several computational and biological domains: foundation models and self-supervised learning, spatial omics and spatial profiling, tumour microenvironment modelling, multiplexed in vivo CRISPR perturbation screens and computational oncology. Rather than relying on public datasets, Noetik generates purpose-built biological data from spatial omics and CRISPR experiments and connects it to its models. The resulting proprietary collection pairs deep spatial profiling with routine clinical assays.
Three platforms sit at the core of the work. OCTO is an AI platform and foundation model for representing and simulating the complexity of human tumours and tissue microenvironments. OCTO-VC is directed at target discovery, biomarker development and prediction of therapeutic response in oncology. Perturb-Map draws on the company's spatial omics and multiplexed in vivo CRISPR data for target discovery and response prediction.
The company works with pharmaceutical and biotechnology partners and frames its mission around translating findings to patients - for example by enriching clinical trials for likely responders and by seeking cancer targets that hold up in people.
