Saptarshi Mondal

SENTRIX : See the cascade before it happens. What if a flooded road could trigger a city-wide emergency? A blocked route can delay ambulances, disrupt access to critical facilities, compromise utility dependencies, and turn a localized incident into a cascading infrastructure failure. Cities don't experience infrastructure failures in isolation. Their roads, utilities, emergency services, and critical facilities operate as interconnected systems. Yet most incident-response tools focus on individual events rather than understanding how one failure can trigger another. SENTRIX is an AI-powered infrastructure resilience intelligence platform built to change that. It transforms fragmented incident information into an actionable understanding of infrastructure risk through four integrated capabilities: 1. Understand the incident. Meta's Muse Spark Open interprets natural-language incident reports, extracts relevant entities, and generates structured hypotheses about potential infrastructure disruptions. 2. Model the city. A graph-based digital twin represents interconnected infrastructure assets and their dependencies, allowing SENTRIX to reason about how disruptions can propagate beyond the initial incident. 3. Simulate the cascade. A deterministic simulation engine models infrastructure state transitions and cascading failures, making potential consequences reproducible and independently verifiable. 4. Compare interventions. Counterfactual scenario analysis enables decision-makers to compare baseline conditions against mitigation strategies, understand changes in disruption and response delays, and identify potentially effective interventions. The critical architectural decision is the separation of AI-generated hypotheses from verified infrastructure data and deterministic simulation results. Model outputs are validated against known digital-twin assets; the simulation engine, not the language model, calculates operational impact. SENTRIX brings together AI-assisted incident understanding, graph-based systems modeling, deterministic simulation, and evidence-grounded decision support in one operational workflow. Our vision: Move cities from reacting to infrastructure failures toward anticipating their consequences. Because resilience is not just about knowing what failed. It is about understanding what could fail next — and acting before the cascade begins.