Data engineering is essentially the plumbing that makes data usable — building and maintaining the systems that collect, move, clean, and store data so that analysts, scientists, and tools can actually work with it.

A few key thoughts:

Core focus areasPipelines (ETL/ELT) – Extracting data from sources (databases, APIs, files), transforming it, and loading it into a warehouse or lake.

Data modeling – Designing how data is structured for both efficiency and usability (star schemas, normalization, etc.).

Infrastructure – Managing databases, warehouses (Snowflake, BigQuery, Redshift), and orchestration tools (Airflow, dbt, Dagster).

Data quality & governance – Ensuring data is accurate, consistent, and well-documented — arguably the most underrated part of the job.

Why it mattersGood data engineering is invisible when done well — analysts get clean, reliable data without needing to know how it got there.

Bad data engineering shows up everywhere: broken dashboards, inconsistent numbers across reports, analysts spending 80% of their time cleaning data instead of analyzing it.Skills in demand right nowSQL (non-negotiable, still the backbone)Python (for pipeline logic, automation)Cloud platforms (AWS/GCP/Azure)Tools like dbt (transformation), Airflow (orchestration), Spark (large-scale processing)Connection to M&E/data analysis workIf you're coming from an M&E or research background, data engineering is a natural adjacent skill — a lot of your work (turning messy field data into clean, analyzable datasets) is essentially light data engineering already.

Learning SQL well and picking up basic pipeline concepts (even just automating a repeatable data-cleaning process in Python) could meaningfully upgrade the kind of consulting/freelance work you can offer.