Fluid Learning Architecture (FLA): a modality-agnostic learning-systems architecture for professional and workplace education, grounded in learning science

Learning systems often appear complete on paper—programmes launched, platforms deployed, curricula defined—yet capability in practice remains fragile, uneven, or short-lived. This preprint introduces Fluid Learning Architecture (FLA), a modality-agnostic learning operating architecture designed around human learning psychology and the socio-technical conditions of transfer. The paper is a conceptual synthesis and design-architecture contribution, and it proposes a practice-based evidence documentation structure to support disciplined iteration across deployments.FLA is structured in two layers. Layer 1 (FLA Core) comprises eight stable rules governing context sensing, intent definition (including experience intent), barrier diagnosis, capability modelling, scaffolded experience design, orchestration and prepared environments, activation of transfer through participation systems, and evidence-led evolution. Layer 2 (Pattern Library) comprises selectable implementation patterns (e.g., transfer disciplines, inclusive design frameworks, simulation and coaching patterns) that can be combined without altering the core. A conceptual foundations section maps the eight rules to established empirical and theoretical traditions spanning learner-centred and experiential approaches, self-determination theory, durable learning and cognitive load, situated/social learning, psychological safety, human factors and safety science, implementation/realist evaluation, and learning analytics.To make practice-based evidence legible and cumulative, the paper proposes “evidence episodes” as a consistent vignette structure for documenting portability across industries, learner needs (including neurodiversity), and technology stacks. FLA is offered as a framework for building learning systems that produce durable capability and evolve through disciplined iteration rather than retrospective justification.Note: Here, “modality” refers to learning delivery channels (e.g., coaching, workflow support, XR simulation, classroom), not ML input modalities.