Engram: A Framework for the Execution Substrate of Modern AI Models

Modern AI research typically treats trained models as mathematical abstractions defined by parameters, architecture, and optimization trajectories. Yet during inference, these abstractions are realized through physical computation on heterogeneous hardware and dynamic runtimes. This paper argues that the behavior of a model during inference-the *realized model*-is systematically shaped by execution-level factors such as precision formats, memory hierarchy, kernel scheduling, numerical stability, and hardware topology. We introduce Engram as a conceptual framework for understanding this execution substrate and its influence on reasoning, coherence, and long-range behavior. Engram reframes inference as the physical simulation of a trained dynamical system under resource constraints, exposing a reproducibility gap and motivating a new scientific focus on execution fidelity. We outline falsifiable predictions, propose a minimal experimental protocol for measuring execution drift, and situate Engram relative to numerical analysis, MLSys, and cognitive AI.