01
IoT Lakehouse Pipeline
A production-inspired reference pipeline for streaming and batch IoT analytics.
Streaming telemetry reaches Delta Lake through Kafka and Spark. DuckDB reads the tables for data-quality reports.
Select a stage to keep its context. Press Escape to clear it.
Problem
Telemetry needs reliable ingestion, quarantine rules, repeatable writes, and useful analytics models.
Role
I built the end-to-end reference implementation.
System
- Kafka ingestion
- Spark Structured Streaming
- Delta Lake Bronze, Silver, and Gold layers
- Idempotent MERGE writes and quarantine checks
- DuckDB data-quality reports
Evidence
Streaming and batch paths share one lakehouse model. Resume source
Silver processing validates and quarantines bad records. Resume source
Next run
Add repeatable failure drills and publish measured recovery behavior.

