A quiet growth story about practice, patience, and people who wanted to understand their work
Some milestones arrive with noise. This one arrived quietly, on an ordinary evening, when the counter moved past one hundred thousand.
One hundred thousand data professionals have now learned something here. Engineers, analysts, students, career switchers, people studying at 5 in the morning before work and people practicing at midnight after their kids fell asleep. People in cities we have never visited, working on problems we will never see.
Not long ago that number was forty thousand. This is the story of the distance between those two numbers.
BricksNotes started with one belief and no audience.
The belief was simple. Most people do not struggle with data engineering because it is too hard. They struggle because it is taught backwards. Tools first, concepts later, practice never. You finish a course, you can repeat the words, and then a real pipeline breaks and you realise you never learned how the system behaves.
So we wrote the opposite of a course. We wrote a book you can run.
Thinking in Data Engineering with Databricks began as a set of lessons meant to be opened next to a notebook, not read on a sofa. Every chapter had something to type. Every concept had something to observe. Databricks Free Edition was the classroom, because a free classroom is the only kind that is fair.
The first readers were a handful of people. Then a few hundred. Then forty thousand.
Growth is useful mostly because it is honest feedback at scale. When forty thousand people move through your content, you stop guessing what confuses them.
Three things became obvious.
People wanted to practice, not just read. People wanted proof they were making progress. And people arriving with zero background needed a first win before they needed depth.
Everything we built after that came from those three sentences.
This is the part that took the time. Not marketing. Building.
Lessons that behave like a lab. The core book grew into a full path across workspace essentials, data sources, DataFrames, Spark SQL, transformations, joins and aggregations, Delta Lake, schema evolution, partitioning and performance, streaming, incremental processing, SCD patterns, testing, debugging, medallion architecture, Unity Catalog, and machine learning basics. The first three lessons stayed free, because nobody should have to pay to find out whether a teaching style fits them.
Practice exams that respect the exam. We built full practice sets for Data Engineer Associate, Data Engineer Professional, and Data Analyst Associate, with explanations that teach the concept rather than confirm the letter. More than twenty five thousand practice exams have been completed. That number matters more to us than page views, because it means people sat down and did the work.
A first project anyone can finish. Your first data engineering project in Databricks Free Edition is a full guided build with messy sample data, bronze to gold, a MERGE, a rerun that does not double count, and honest troubleshooting. Almost everyone breaks it on the first run. That is the point. The break is where the learning is.
Reference material for the working day. PySpark cheat sheet, SQL cheat sheet, unit testing cheat sheet, datasets you can load in one line, and explainers for the questions beginners actually type into a search box: what Databricks is, medallion architecture, Jobs and Pipelines, Auto Loader, Unity Catalog, pricing.
Tools instead of tutorials. The PySpark pipeline generator writes a production-shaped pipeline from a short form, then shows you why each choice was made. The resume builder and interview readiness tracks exist because learning is only half of a career change.
Proof of progress. Chapter quizzes, badges, and verifiable certificates came directly from that second lesson at forty thousand. People do not want gamification. They want evidence.
Field notes, written daily. The blog turned into a practice of its own. Late-arriving data. Safe backfills. Schema drift. Streaming tables against materialized views. Liquid clustering against Z-ORDER. Data quality checks that catch the failures nobody gets paged for. Over a hundred articles now, each one written the same way a good colleague explains something at a desk.
A wider view of the industry. BricksNotes Intelligence covers the Data and AI world beyond one platform, and BricksNotes Watch gathers the videos for people who learn better by listening.
And a second book, for what is coming next. As agents started reading catalogs and writing to tables, the skill underneath the tools changed. That work became The Context Advantage, about the discipline of giving systems the right context. Our article on context engineering as a real job skill is the short version.
We wish there was a clever answer. There is not.
We published on the days it felt pointless. Consistency beats brilliance over a year. Most articles did nothing on the day they went out and quietly kept finding readers for months.
We answered real questions. Not keyword lists. The actual sentence a tired engineer types at 11pm when a job ran twice and the numbers doubled.
We refused hype. No ultimate guides, no guaranteed outcomes, no fear based urgency. Trust compounds slower than clicks and lasts much longer.
We kept the free part genuinely useful. A large share of what we have built costs nothing. That is not a funnel trick. It is the point of the project.
We listened when people told us something was broken. Half of the improvements in the last year started as one sentence from one reader.
One hundred thousand is an abstraction. The real story is smaller and better.
It is the analyst who wrote her first PySpark job and did not need help. The engineer who passed the Associate exam on his second attempt after failing the first and telling us so. The student in a town with no data team who now understands medallion architecture better than some people who get paid for it. The person who changed careers at 38 and told us they finally felt like an engineer.
That is what a hundred thousand actually is. A hundred thousand ordinary evenings of effort.
We are not treating this as an arrival. Here is what we are working toward.
More guided projects, so people finish things rather than collect notes. Deeper coverage of AI engineering on the lakehouse, because that is now part of the job and not a separate career. More practice, better explanations, and content that stays current as the platform moves.
And one goal that sounds large because it is. We want any data professional anywhere with an internet connection to be able to learn this work properly, without paywalls on the fundamentals and without needing a company to sponsor their curiosity. The world is connected now. Access to good teaching should be too.
If you are one of the hundred thousand, thank you. You built this by showing up.
If you are new, start where everyone starts.
Run it. Break it. Run it again. That has always been the whole method.