Mission
We build computational models to analyze the immune system and make it interpretable.
What we solve
Our work focuses on deep learning for immune single-cell data, with an emphasis on quantifying uncertainty and understanding immune dynamics over time.
Our flagship model annotates immune cells at fine resolution and quantifies confidence across every prediction.
Methodology
Our models are developed in close collaboration with immunology and single-cell domain experts from UCSF, with rigorous oversight at every stage.
This includes dataset curation, immune cell definitions, benchmarking, and systematic evaluation of model confidence and failure modes.
Get in touch
We're always interested in hearing from the research and biotech community. Reach out if you have questions about our tools, would like to discuss our methods, or are interested in collaborating.
People and affiliations
Led by Amadeo and Al, in collaboration with researchers from the UCSF Data Science Colab, with additional support from the UCSF Data Library.