What learners say after doing the work
Reviews from engineers who completed the Python, Transformer, and Capstone courses. Dates, roles, and details are as submitted.
Back to HomeLearner reviews
I had been working with pandas for two years but genuinely did not understand memory layout or why my pipelines were slow. The Python course fixed that. The dataset audit on the open data portal was the most useful exercise — it took me several tries to get the environment reproducible, which was the point.
The Transformer course is not easy. I came in comfortable with PyTorch and still found weeks four through six hard. The paper-reading sessions helped, but the real work was in the assignments — building attention from scratch and watching it fail the first three times before it worked. Code review feedback was specific and honest.
I enrolled in the Capstone Track after finishing the Transformer course. The mentor I was assigned to had clearly read my previous work before our first session, which set a useful tone. The architecture reviews were the most useful thing — having a second reviewer push back on my retrieval design decisions was something I had never experienced in a course before.
The Python course covered things I thought I already knew, but the memory profiling week showed me I did not. The bit that surprised me was how much of it was about failing loudly — catching issues in the pipeline before they show up silently downstream. That framing changed how I write production code generally.
The dataset licensing and consent session in the Transformer course was one I did not expect to find useful, but it changed how I think about training data in my day job. The public write-up at the end felt like a lot of work at the time. Looking back it was worth the effort because it forced me to explain what I actually understood versus what I had just copied.
The incident post-mortem in the Capstone Track was the most unusual assignment I have ever done in a course. Writing it properly required me to be honest about a design decision I made in week nine that caused problems in week eighteen. The written evaluation methodology also getting external critique meant I could not just describe what I wanted to do — I had to defend it.
Case studies
Data analyst moving to model work
An analyst at a logistics company in KL had been working with structured data for four years but found that every model training tutorial assumed Python knowledge that went beyond her pandas work. Notebooks ran but pipelines failed silently in production.
Enrolled in the Python and Data Engineering course. Spent weeks five and six on columnar formats and week seven on reproducible environments. The environment reproducibility check on every submission forced a habit of explicit dependency declaration that she had not practised before.
Completed the ten-week course with nine assignments reviewed. Moved into a data engineering role within the same organisation three months after close. Went on to enrol in the Transformer course.
"The dataset audit on Malaysian open data was the right level of hard. Not artificial."
Backend engineer wanting to implement from paper
A backend developer had tried three different online transformer courses. Each one used a high-level library that abstracted away the mechanics he wanted to understand. He could call the API but not explain what it was doing.
Enrolled in the Transformer Architectures course. Worked through the Attention Is All You Need paper in the reading sessions and implemented each component as an assignment. The shared cluster meant he could run training jobs that took hours without needing his own hardware.
Eighteen weeks, sixteen assignments reviewed. Produced a small transformer trained on a public Malay language corpus. Published a write-up documenting what worked and what the failure modes were.
"I finally understand why my loss was not going down. That took until week eleven."
Contact
KL Gateway, 59200 KL
Mon–Fri 9AM–6PM
Sat 10AM–2PM
Read the course specifications and then enquire
The solutions page has the full curriculum. The contact form is on the homepage.