Due to the highly controlled nature of clinical trials, it is difficult to capture much-needed data about patient variability across a condition that translates into the real world. This is particularly problematic for immune-mediated inflammatory diseases since individual patient responses to a therapy can vary significantly. To address this challenge, we have built a proprietary machine learning platform which links clinical notes, lab samples, and biopsies to model immuno-inflammation pathways at an individual level. By applying our platform to curated datasets, we create predictive models for personalized patient diagnosis, prognosis, and response to treatment. Our model for rheumatoid arthritis treatment was developed in partnership with key opinion leaders and drew from patient records from four separate European databases. Our results have been presented at scientific congresses and published in a peer-reviewed journal. We will be modeling 11 additional auto-immune diseases drawing on access to approximately 5 million medical records of patients with auto-immune diseases through a strategic partnership with the French Parisian Hospitals group.