IoT Telemtr Pipeline Course — DevOps skill for Claude Code

IoT telemetry pipeline — build guide

[](https://github.com/Bantosito/Iot-Telemtr-Pipeline-Course/actions/workflows/ci.yml)

A sensor at home streams readings into Azure, where they're stored, modelled and shown in a report. The cloud side is defined in code and deployed automatically.

This file is the plan. **It contains no code on purpose.** Each step tells you what the piece is, why it exists, what to build, how you know it's done, and questions you should be able to answer afterwards.

**How this was built:** I built this project with an AI pair programmer (Claude Code), which wrote much of the code with me. To make sure I understood every piece rather than just collecting it, each step has a gate: before moving on, I answer that step's questions in my own words in [`NOTES.md`](NOTES.md), my learning log for the whole project.

 AT HOME                                          AZURE
 ┌──────────┐  MQTT  ┌───────────┐  ┌────────┐     ┌─────────┐  ┌──────────┐  ┌────────────┐  ┌─────────────┐
 │  sensor  │ ─────► │  broker   │─►│ bridge │ ──► │ IoT Hub │─►│ Function │─►│ PostgreSQL │─►│ dbt →       │
 │ (fake →  │        │(Mosquitto)│  └───┬────┘     └─────────┘  └──────────┘  └────────────┘  │ Power BI    │
 │  ESP32)  │        └───────────┘      │ bad readings                                         └─────────────┘
 └──────────┘                           ▼
                                  dead-letter topic

---

How to use this guide

• One step at a time, in order. Each builds on the last.
• Commit at the end of every step, with a message saying what now works. Your commit history is part of the portfolio.
• Answer the questions in NOTES.md in your own words. If you can't answer one, you haven't finished the step. These answers are your interview preparation.
• Stuck for 45 minutes? Stop. Write down exactly what you expected, what h