An 18-month engineering journey building an event detection system at the edge using Azure IoT Operations (AIO). The team evolved from jq-based Data Processor Pipelines to a hybrid architecture combining custom Rust pods with AIO Dataflows (expression-based and WASM-based). Rust pods handle heavy transformations, enrichment, and heuristic logic, while Dataflows manage routing, filtering, and cloud connectivity. MQTT serves as the central nervous system, enabling decoupling, tracing, and horizontal scalability. Key lessons include favoring modularity, embracing hybrid approaches over full rewrites, and treating the MQTT broker as the architectural backbone for edge AI systems.
Table of contents
The Evolution: From jq Pipelines to WASM Dataflows Copy linkHybrid Approach: Rust Pods + Dataflows Copy linkA Deep Dive Into Our Transformations Copy linkScaling the Architecture Copy linkMQTT as the Brain Copy linkLessons Learned Copy linkThe Takeaway: Design for Change Copy linkSort: