Sapan Pandya

Designing Trustworthy AI Driven Decisions in IoT Systems

Abstract:

 The Internet of Things (IoT) systems increasingly depend on AI driven decision logic deployed on devices and edge nodes. In high risk software systems, unsafe or opaque decisions can cause financial loss, operational failures, or compliance violations. This presentation focuses on designing trustworthy AI driven decision components for IoT systems. It presents concrete design patterns for running inference under resource constraints, validating AI outputs in real time, and enforcing safe state transitions. The talk highlights engineering techniques such as deterministic pre checks, confidence based decision gating, invariant validation, and audit ready logging. Drawing from real world large scale deployments, the presentation demonstrates how trust, predictability, and regulatory alignment can be built into AI enabled IoT systems without sacrificing performance or autonomy.  

Profile:

Sapan Pandya is a Senior Software Engineer from the United States with more than a decade of experience in designing, developing, and deploying large-scale, high-performance technology platforms. His expertise spans full-stack development, system architecture, microservices, and edge-computing solutions supporting complex, regulated environments.
He has led end-to-end engineering for lottery and retail sports betting platforms across multiple U.S. states, focusing on secure transaction systems, modular architecture, and reliable field-deployed terminals. His professional interests include system scalability, software reliability, and sustainable technology practices. In addition to his industry work, he serves as a reviewer for IEEE, Springer Nature, and several other international journals and contributes to the program committees of global technology conferences.