LLM assistant for IoT Platform

Abstract

This thesis explores a novel approach to this problem by leveraging advances in Large Language Models (LLMs) as an intelligent orchestration and configuration layer. The core deliverable is a program that accepts a structured or free-form user description of an intended IoT solution (for example: target devices, data flows, latency and security constraints, analytics needs) and automatically generates the necessary configuration, infrastructure orchestration, and deployment actions to instantiate a customized IoT platform. The system combines LLM-based requirement interpretation with deterministic infrastructure tooling (infrastructure as code, container orchestration, device provisioning scripts, and CI/CD pipelines), applying policy templates and security best practices where appropriate. The result is an end-to-end path from natural language intent to a functioning, observable IoT deployment.