ORBE
“We advance the frontiers of artificial intelligence for medicine and healthcare. Our research spans deep learning, multimodal AI, reinforcement learning, and autonomous agents to build intelligent systems that learn, reason, and improve clinical decision-making.”
- Medical Imaging
Deep learning for neuroimaging
Segmentation and analysis of brain MRI — from white-matter lesions to full anatomical structures — built to hold up on real clinical data, not just curated benchmarks.
- Robustness
Domain adaptation and generalization
Consistency and adversarial training that keep imaging models reliable when the scanner, site, protocol, or population changes underneath them.
- Trustworthy AI
Uncertainty and calibration
Probabilistic augmentation and calibrated predictions, so a model can say how confident it is before a clinician acts on its output.
- Multimodal AI
Learning across images, text, and records
Models that read imaging alongside clinical notes and structured records, reasoning over the whole patient rather than a single modality.
- Reinforcement Learning
Learning to decide
Decision policies that improve with experience, applied to care pathways, triage, and resource allocation in constrained health systems.
- Autonomous Agents
Agents for clinical work
Agentic systems that read documents, verify evidence, and draft reports — the research line behind the ORBE Health platforms.
Mauricio Orbes, PhD
Founder · Research Scientist
Machine learning researcher with over eight years of combined academic and industry experience applying AI to healthcare. He has worked with cross-functional teams to identify data-centric challenges and deliver intelligent solutions in radiology, teledermatology, and decentralized clinical trials. His research spans transformer-based vision models, reinforcement learning, and domain adaptation for robust and interpretable medical-AI systems — the foundations of ORBE’s work on data-driven technologies that improve patient care and outcomes.
Research inquiries — contact@orbelabs.ai