// testimonials
What participants say
after studying with us
Written by engineers and analysts who completed one or more Tensorhaus courses. Unedited except for minor formatting.
Back to Home3+
Years running courses
180+
Practitioners enrolled
4.7
Average rating from 5
96%
Lab completion rate
// reviews
Participant Reviews
Wei Zhen
Data Engineer · Petaling Jaya
"I had been writing pipeline code for about four years but had no clear picture of how a trained model fits into that kind of infrastructure. The MLOps course filled that gap specifically. The experiment tracking module was the most useful part — I have already applied it at work."
MLOps Fundamentals · April 2025
Siti Rahayu
Business Analyst · Kuala Lumpur
"I work with sales data and had been using Excel models. The Tabular Data course taught me how to approach the same problems with gradient-boosted trees in a way that was methodical rather than trial-and-error. The cross-validation section was particularly clear. I would have appreciated one more week on feature engineering."
Tabular Data Methods · March 2025
Khairul Hisham
Backend Engineer · Shah Alam
"The Friday session suits my schedule well — one hour on a Friday afternoon, no homework. What I find useful is hearing Kian Wei reason through code he is reading for the first time. The reasoning is visible in a way it is not when someone explains polished code."
Friday Code Reading · Ongoing since February 2025
Lee Mei Shan
ML Engineer · Subang Jaya
"I took the MLOps course after about a year of running models ad-hoc in notebooks. The section on production monitoring was genuinely new to me — I had not thought systematically about data drift before. The lab setup was solid; I did not spend time fighting dependency issues, which I had expected to."
MLOps Fundamentals · January 2025
Fadzillah Ismail
Data Analyst · Klang Valley
"The Tabular Data course is compact but covers the important parts. I had used random forests before without understanding why certain validation practices matter. The course made those reasons clear. The notebooks are clean and worth keeping as reference material."
Tabular Data Methods · April 2025
Rajesh Thambipillai
Senior Software Engineer · Kuala Lumpur
"I subscribe to the Friday session primarily to maintain exposure to ML code outside my current project. It takes exactly one hour of my week. The choice of code is usually interesting — I have seen LightGBM internals, a simple recommender, and a data preprocessing utility. The commentary is the value; reading alone is slower."
Friday Code Reading · Ongoing since December 2024
// case studies
Participant Journeys
// case 01 · MLOps Fundamentals
Challenge
A data engineer at a logistics company was responsible for running a demand forecasting model. It lived in a notebook, re-run manually each week, with no tracking of which parameters had been used or how output had changed over time. When results looked different, there was no record to investigate.
How the Course Helped
The MLOps Fundamentals course introduced experiment tracking with MLflow and a systematic approach to packaging the model as a reproducible artefact. The participant rebuilt the forecasting workflow over the ten weeks, converting it from a manual notebook into a versioned, tracked pipeline.
Result
The rebuilt pipeline logs every run, stores the artefact version used, and flags when model performance drops below a defined threshold. Debugging now takes minutes rather than hours. The participant's team adopted the same pattern for two other models.
"The course gave me a vocabulary and a set of habits I could immediately apply. The ten-week pace was sustainable with a full workload." — Wei Zhen, Data Engineer
// case 02 · Tabular Data Methods
Challenge
An analyst working in retail planning had been building regression models to estimate sales uplift from promotions. The models were linear and trained on a split chosen arbitrarily. Results were unreliable and the analyst was not confident in the methodology.
How the Course Helped
The Tabular Data Methods course addressed cross-validation methodology directly, including how to avoid leakage and how to interpret evaluation metrics on held-out data. The participant rebuilt the promotion model using LightGBM with proper stratified cross-validation.
Result
The revised model holds up better on new promotions. More practically, the analyst now understands why it does or does not — the feature importance output is readable and the validation procedure is defensible when explaining results to stakeholders.
"I came in thinking I knew enough about trees. I did not. The course was compact but honest about what matters." — Siti Rahayu, Business Analyst
// reach us
Contact Tensorhaus
+60 3-4257 9083
Monday – Friday, 9 am – 6 pm
For enquiries and enrolment
Suite 6.3, Plaza Ampang City
Jalan Ampang, 50450 Kuala Lumpur
Office Hours
Mon – Fri: 9:00 am – 6:00 pm
Saturday: 10:00 am – 2:00 pm
// enrol
See if a course is right for you
Send us your current role and which course interests you. We will answer your questions about prerequisites and structure before you enrol.
Send an Enquiry