The ACS platform is architected on AWS as a robust SaaS solution, designed for scale, resilience, and high-frequency causal reasoning at the edge.
Core AWS infrastructure diagram showing the interaction of edge sensing, cloud-native services, and the causal reasoning engine.
This edge component handles high-frequency raw telemetry. It is crucial for security and performance, performing data pre-processing and secure ingestion near the source. In our pilot, it extracts physics-aware features from mmWave radar and pushes them through sidecar connectors for private, real-time causal deconvolution.
This layer provides the fundamental **Speed** and **Thematic Reasoning** capability. It's not a generic vector database; it's designed specifically for causal geometry. It stores high-dimensional representations of domain concepts, trained by a **Human-in-the-Loop** to prioritize mathematically valid causal links. Crucially, it retains a persistent **Historical Memory of Decisions**, using past interventional outcomes to refine the current reasoning logic continuously.
This visual component is the source of truth for the **Aura Causal Manifest (ACM)**. These human-readable, auditable DAGs (like the one below) map the structure of world mechanisms. The DAG isn't just a visualization; it's the executable logic for Pearl’s do-calculus. It ensures that every institutional decision can be scientifically justified through counterfactual simulation, reducing reliance on correlation-based black-box models.