At this point, everything was finally coming together.
- The learning application had been adapted for certification preparation.
- The card generation workflow was working.
- The decks were growing steadily.
- My study routine was becoming consistent.
Most importantly, I had a plan.
A comfortable plan.
My original goal was simple.
I wanted to take my time.
Not rush.
Not memorize.
Not chase an exam date.
I wanted to understand the material properly.
Explore concepts.
Experiment with Azure services.
Build confidence gradually.
And sometime toward the end of the summer, I would schedule the exam.
At least, that was the plan.
Then one sentence changed everything.
During a regular team meeting, one of my colleagues mentioned that Microsoft was retiring several certifications and replacing them with newer alternatives.
At first, I didn’t think much about it.
Microsoft updates certification paths all the time.
It sounded like one of those announcements that mattered to somebody else.
Then I looked at the list.
And suddenly it mattered to me.
A lot.
The certification I had chosen was on it.
I remember staring at the announcement and trying to process what it meant.
Surely I had misunderstood something.
Surely there would be plenty of time.
There wasn’t.
After checking the official information, the situation became very clear.
The certification wasn’t disappearing tomorrow.
But it was disappearing soon. Much sooner than I had expected.
And suddenly the timeline I had carefully built for myself no longer existed.
Instead of several relaxed months of preparation, I had roughly six to eight weeks.
Maybe a little more. Maybe a little less.
Either way, the difference was dramatic.
The challenge wasn’t Machine Learning itself.
I already had previous experience.
I had completed courses.
Built projects.
Worked through exercises.
The challenge was Azure.
Most of my Azure Machine Learning exposure came from a guided workshop environment.
Valuable experience. But limited.
And certainly not enough to walk confidently into a certification exam.
So I made a decision.
A very simple one.
I would attempt it.
Not because I was certain I would pass.
I wasn’t.
Not because I felt ready.
I definitely wasn’t.
But because I wanted to see how far focused effort could take me.
For a few days, I considered the obvious option.
Postpone. Skip it. Accept that I had started too late. Wait for the replacement certification.
Take the longer route.
There would have been nothing wrong with that decision.
But the more I thought about it, the more I realized something.
I had already built the system.
I had already committed to the journey.
I had already invested countless hours creating study materials.
And deep down, I knew I would regret not trying.
That decision changed the entire nature of the project.
Up until that point, I had been studying out of curiosity.
Now I had a target.
A date.
A deadline.
And suddenly every study session mattered.
My learning strategy had to change.
The exploratory approach that I normally enjoy was no longer enough. There wasn’t time to investigate every interesting topic. There wasn’t time to follow every rabbit hole.
I needed focus. Prioritization. Discipline.
For the first time during the project, I wasn’t just learning.
I was preparing.
- The exam date went into the calendar.
- The countdown began.
And from that moment forward, the pace changed completely.
- Evenings.
- Weekends.
- Travel time.
- Lunch breaks.
Every available hour started moving toward a single goal.
Looking back, this was probably the moment when the project stopped being a personal learning experiment.
And became a mission.
What I didn’t realize at the time was that the certification deadline would not be the only unexpected challenge waiting for me.
A few weeks later, another problem appeared.
This one had nothing to do with Microsoft.
Nothing to do with Machine Learning.
And everything to do with a bus, an iPad, and the mountains of Transylvania.
In the next chapter, I’ll tell the story of how an upcoming trip forced me to redesign parts of my learning system and build an offline version of an application that was never supposed to work without an internet connection.