// learning.hub

Learn data engineering by shipping real systems.

Every track has a duration, prerequisites, and a final project you can put in front of a hiring manager. No filler.

// roadmap · beginner → production engineer5 stages

01Beginner
Foundations
→
02Beginner
Data Ingestion
→
03Intermediate
Analytics Engineering
→
04Intermediate
Streaming
→
05Advanced
Platform Engineering

Foundations

Git · Linux · Python · SQL

live

GitBashPython 3.12PostgreSQL

duration: 3 weeks
difficulty: Beginner
prereqs: None

outcomes

→Comfortable in a terminal
→Read/write basic SQL
→Ship a Python script

final project

Build a CLI that queries a Postgres DB and commits results to a git repo.

Start track →repo ↗

Data Ingestion

dlt Fundamentals Workshop

live

dltREST APIsParquetDuckDB

duration: 2 weeks
difficulty: Beginner
prereqs: Python basics, SQL basics

outcomes

→Extract from any REST API
→Load to warehouse
→Handle incremental loads

final project

Ingest OpenF1 telemetry into DuckDB with incremental cursors.

Start track →repo ↗

Analytics Engineering

dbt · DuckDB · Warehousing

beta

dbtDuckDBSQLJinja

duration: 4 weeks
difficulty: Intermediate
prereqs: SQL, Basic modeling

outcomes

→Staging → marts
→Tests + docs
→Exposures to BI

final project

Model the F1 dataset into a star schema with tests + exposures.

Start track →repo ↗

Streaming

Apache Kafka Learning Track

live

KafkaSchema RegistryKafka Connect

duration: 5 weeks
difficulty: Intermediate
prereqs: Docker, Python or Java

outcomes

→Design partitioned topics
→Write idempotent producers
→Reason about rebalancing

final project

Ship a producer/consumer pair that survives a broker restart.

Start track →repo ↗

Flink + Debezium — Stream Processing

planned

FlinkDebeziumCDC

duration: 4 weeks
difficulty: Advanced
prereqs: Kafka track

outcomes

→CDC pipelines
→CEP rules
→Windowed aggregations

final project

Detect velocity fraud on a live Postgres → Kafka → Flink pipeline.

Start track →repo ↗

Platform Engineering

Docker · Kubernetes · CI/CD

planned

DockerKubernetesGitHub Actions

duration: 4 weeks
difficulty: Intermediate
prereqs: Comfortable with the terminal

outcomes

→Containerize services
→Deploy to k8s
→Automate tests + releases

final project

Package the fraud platform's services into a working k8s deployment.

Start track →repo ↗