Designed and shipped an automated hardware validation pipeline for IoT devices: a Node service drives device commands over MQTT, validates sensor/telemetry responses against expected state, and streams live results over WebSockets to a React console cutting post-install QA from manual checklists to a single automated run.

Made the production backend hardware version aware so a new device generation could ship on the same APIs, schedulers, and alerting paths model specific control payloads, lifecycle scoring, and guards so old devices kept working while new ones came online.

Debugged and fixed production reliability issues in a high-noise IoT environment: duplicate push/WhatsApp alerts, false-positive service tickets from flaky sensor bursts, and job-queue floods that were cascading retries under load.

Built the device onboarding path end-to-end (inventory → customer bind → telemetry activation) behind internal APIs and an admin UI, so provisioning stopped being a multi-step ops/DB workflow

Shipped changes across the customer-facing PWA, field-ops PWA, and core API that run the live connected-device product.

Tech Stack: Node.js, Next.js, PostgreSQL/TimescaleDB, PgBoss, MQTT, EMQX, Grafana, Jenkins/PM2, Websockets