AWS Certifications

AWS
Published

2026/09/26

I set myself a few goals for 2026. These were the big ones:

The preparation for the ultras went well, despite a dodgy knee, and I finished both races earlier in the year. I’m still grinding away at those books. At the rate I’m going I doubt I’ll hit that target.

I’ve chipped away at the AWS preparation on train trips and during lunch hours, listening to lectures, taking notes and doing mock exams. I took the fourth and final exam yesterday.

When I woke up this morning I saw an email from AWS. But I’m getting ahead of myself. Back to that later.

Certified Cloud Practitioner — Foundational (CLF-C01)

I started with Certified Cloud Practitioner back in January 2023. I’d already been dabbling in AWS for a few years, and a Foundational certification seemed a reasonable place to start.

I spent some time preparing for the exam. It wasn’t too gruelling: 65 questions in 90 minutes (just under 90 seconds per question), with a pass mark of 700 out of 1000.

I decided to sit the exam at home. I hadn’t anticipated quite how intrusive that would be. I couldn’t use my Linux machine, so I had to dredge up an old Windows laptop. Then I had to practically clear a room in the house to satisfy the online proctor. What a mission! I passed fairly comfortably, but decided that any further exams would be at an exam centre.

I recertified earlier this year. The version I originally sat, CLF-C01, has been replaced by CLF-C02. The changes seem to centre on the weighting of the four domains, with some updates to the content within them.

Certified Data Engineer – Associate (DEA-C01)

I kicked off this year’s certification campaign with Certified Data Engineer. Moving up to Associate meant a big step up in scope and depth. It took me about six weeks to get ready.

I prepared with a course and practice exams on Udemy. This was my approach:

  1. Watched the course videos and took notes on paper.
  2. Did all of the practice exams.
  3. Revisited the topics behind questions I’d got wrong or found tricky, and updated my paper notes.
  4. Took a few of the exams again. My results improved enough for me to feel ready.
  5. Summarised my notes and used them to brush up on the morning of the exam.

There were 65 questions again, but this time I had 130 minutes (two minutes per question), with a pass mark of 720 out of 1000. More time. Higher pass mark. The extra time soon made sense: many of the questions were long and involved, with lots of details to keep track of. It could take around two minutes just to read a question and its possible answers, never mind making a decision. Keeping up with the clock was a challenge.

I sat the exam in March and passed by a reasonable margin.

Certified Solutions Architect – Associate (SAA-C03)

Buoyed by that pass, I got straight on with Certified Solutions Architect. There was a lot of overlap with the Data Engineer material, so preparation went quicker. I was ready in around four weeks.

I used a course and practice exams on Udemy again. My approach was similar, but I started taking notes in Obsidian. I also skimmed some of the videos and spent more time on the practice exams.

The exam structure was the same as for Data Engineer. I sat it in May, seven weeks after DEA-C01, and this time passed by a more comfortable margin.

Certified Machine Learning Engineer – Associate (MLA-C01)

By this point, a lot of the material was familiar from the two previous Associate certifications. There was still a bunch of new content to work through, most of it revolving around SageMaker. Preparation took around four weeks again.

I used a course and practice exams on Udemy, following the same study routine but shifting even more of my time towards the practice exams.

I sat the exam at the beginning of July, another seven weeks after SAA-C03, and got exactly the same scaled score as Solutions Architect. I wouldn’t read too much into the matching numbers, since they’re different exams, but it was a pleasing coincidence.

📌 The English version of the MLA-C01 exam is only available until 28 September 2026. Beta testing for its replacement, MLA-C02, begins the following day.

Certified Generative AI Developer – Professional (AIP-C01)

Moving up to Professional meant another jump in depth and scope. There was still plenty of overlap with the Associate certifications, but I also needed to cover Bedrock, knowledge bases and agents.

I’d booked an exam slot for the end of September, just over twelve weeks after the MLA-C01 exam, expecting to move it if preparation or other commitments got in the way. That gap wasn’t all productive study time. I was rather burned out and it was summer. Being indoors with lectures and notes had lost most of its appeal.

I’d worked through a course and some practice exams on Udemy, and used Obsidian for revision. I’d also dabbled in Skill Builder, which, in retrospect, deserved more of my time. But I still didn’t feel ready to sit the exam.

A calendar reminder popped up two weeks before the exam. Then another a week before. Still not ready. I tried to reschedule, but there were no slots at my preferred exam centre for the foreseeable future. It looked like I’d booked the last date they were offering this exam. Weird! 😱 Other centres had slots, but getting to them would have been seriously inconvenient.

With six days left, I had little choice but to knuckle down. I concentrated on the practice exams, revisiting the topics that had tripped me up and tweaking my Obsidian notes as I went. That last week evaporated. Suddenly it was exam day.

The exam has 75 questions and lasts 180 minutes: just under two and a half minutes per question. That sounds adequate. Some questions were a veritable wall of text though, and it took me a few minutes just to work out what they were asking. There were plenty of those. Fortunately there were some shorter questions too. Three hours is a marathon exam, but it goes quickly when you feel like you’re running out of time.

The exam also exposed gaps in my preparation. It had more questions that mentioned Lake Formation than I’d expected. The topic had barely appeared in the mock exams. There was also a question that mentioned Amazon Quick, the successor to Amazon Q Business. A lecture on Quick was added to the course after I’d finished the lectures, so it wasn’t even on my radar.

I finished, but the last five questions were done in as many minutes. They were more guesses than answers. I regretted doing all the practice exams in Practice mode instead of the timed Exam mode. A bit of practice with the clock running would have helped. Being accustomed to operating at near-fatal levels of adrenaline and cortisol would have been useful too.

By the end I was mentally preparing myself for failure. And wrestling with the idea of having to do all that again.

But back to that email from AWS:

Congratulations on passing your AWS Certification exam!

What a relief! The report said PASS, although my score showed just how close it had been. I’d squeaked through. The squeakiest of squeaks. I’ll take it though: a pass is a pass.

That completed the four certifications I’d set out to do this year, working through data engineering, architecture, machine learning and finally generative AI development.

An added bonus of passing was that AWS extended the expiration dates of my other certifications (except Certified Solutions Architect) to match this one. Nice.

The difficulty finding another slot also got me thinking about where to sit any future exams. A meeting room at work might be a useful compromise between home and an exam centre. It would be easier to sanitise than a room at home, and the internet connection is probably more reliable too. Worth considering.

Cloud Practitioner

Data Engineer

Solutions Architect

Machine Learning Engineer

Generative AI Developer

Certification Level Questions Duration Cost Passing Score
Cloud Practitioner Foundational 65 90 min $ 100 700/1000
Data Engineer Associate 65 130 min $ 150 720/1000
Solutions Architect Associate 65 130 min $ 150 720/1000
Machine Learning Engineer Associate 65 130 min $ 150 720/1000
Generative AI Developer Professional 75 180 min $ 300 750/1000

I get a warm feeling when I see all of those on my certification dashboard. Will they move the needle in any way? That remains to be seen. What I do know is that I learned an awful lot in the process.