Data Engineer Roadmap

You want to build and operate the pipelines everything else runs on

SQL, modelling, orchestration and streaming - then the ML systems your data feeds.

The route

Week 1 - Data Engineering: The role itself - warehousing, dimensional modelling, orchestration, batch and streaming, quality, and two build-it projects
Week 1 - SQL for Data: The language of the job - joins, aggregation, window functions, CTEs, indexing and query performance
Week 2 - Python for ML: Python for data work - processing, NumPy, pandas, and the idioms that keep a pipeline readable
Week 3 - Git & Version Control: Pipelines are code - branching, review, and the discipline that makes a change reversible
Week 3 - Jupyter & Colab: Notebook environments for exploring a source before you model it
Week 8 - Software Engineering for AI: The engineering half of the job - testing, CI/CD, code review, and project structure for code that runs unattended
Week 8 - Statistics: The maths behind a quality check - distributions, spread, and what makes a threshold defensible
Week 8 - Foundations: Arithmetic and notation used throughout the sizing and cost calculations
Week 9 - MLOps & Deployment: How your pipeline ships - Docker, CI/CD and monitoring, without the model-serving specialism
Week 9 - AI on Cloud: Cloud identity, access and cost guardrails - the two things a data engineer owns in someone else account
Week 10 - ML Algorithms: What your customers do with the data, so you can model for it rather than guess
Week 10 - LLM: What an AI team will ask your pipelines for - embeddings, retrieval, and why chunking is your problem
Week 11 - Interview Q&A: Data engineering loops are standardised - SQL under pressure, modelling, pipeline design, and the green-pipeline-wrong-number debugging round
Week 11 - AI System Design: Pipelines are half of an AI system. This is where the other half gets sized, and where a freshness target stops being a number and becomes an architecture