Gerrit Ian
van Schalkwyk, MD

Developing a translational framework between clinical presentation and genes, cells, circuits, and computations. The model decomposes presentations into capacities, processes, states, representations, and learning/plasticity variables, creating a clinically useful functional layer for reasoning and treatment research without treating DSM labels as mechanistic explanations.

Architect and clinical lead for Dynamic AI Modeling of an INDividual, a longitudinal person-modeling framework that integrates heterogeneous narrative and structured data. Current R&D is simplifying the architecture around strong base-model reasoning, structured context, provenance, explicit uncertainty, targeted clarification, and human review.

Developing an AI-enabled shared model in which encounters and longitudinal data update an evolving representation of a child’s strengths, vulnerabilities, functioning, exposures, and interventions. Clinicians retain responsibility for correction, interpretation, and action.

Developing a practical architecture for source and identity boundaries, separation of documented facts from inference, evidence lineage, uncertainty preservation, verification, stopping rules, and bounded transformation. The accompanying specification-literacy work argues that safe clinical AI requires more than prompt quality.

Building interactive editions for books, academic chapters, and educational content, with text-grounded search and optional AI layers. Prototypes include pediatric clinical neuroscience publishing and a CME model designed to preserve editorial control while making static knowledge more explorable.

Designed and tested an AI-assisted workflow that converts clinical records into standardized review summaries, decision-support questions, and structured documentation while preserving physician determination. AI acts as a controlled transformation layer rather than an autonomous adjudicator.

Concept development for a partnership-based initiative to preserve and represent indigenous African knowledge, values, and cultural frameworks in future AI systems, including Ubuntu-informed perspectives and community-governed collection strategies.