Teaching AI Engineering
the Way Engineers Learn
Tensorloom was founded to address the gap between introductory tutorials and the actual work of building and maintaining AI systems in production.
Back to HomeWhere Tensorloom Started
Tensorloom opened in Cyberjaya in 2021, set up by a small group of engineers who had spent years working on data pipelines, model training and deployment at organisations across Southeast Asia. What they noticed was consistent: developers entering the field had often learned from fragmented online resources, and struggled when they encountered the kind of problems that show up in a real codebase — feature drift, evaluation games, serving latency under load, and the slow erosion of model quality over months of production use.
The idea behind Tensorloom was to build tracks that cover the whole arc — from statistical reasoning through network training to the operational work of keeping a system honest after it ships. The curriculum is structured the way a working engineer would actually move through a problem: foundations first, then the hard parts, then the parts that most tutorials never reach.
The name comes from two ideas that sit together. A tensor is the core data structure of modern AI work — a multidimensional array that carries the numbers through every computation. A loom is an instrument for weaving threads into structure. The school's design system takes this literally: the weekly schedule is drawn as a woven grid of course threads and skill columns, so a learner can read a programme either as a path through time or as a set of capabilities being assembled.
Cyberjaya was a deliberate choice. The city was built to house technology and engineering work, and its infrastructure and community make it a reasonable base for the kind of school Tensorloom set out to be. Learners across Malaysia — and across the region — study online, in the evening, alongside their working days.
People Behind the Tracks
Azlan Rahman
Co-Founder · Curriculum Lead
Machine learning engineer with eleven years across fintech and logistics. Designed the Foundations and Applied Deep Learning tracks and leads code review for both.
Siti Thenaruban
Co-Founder · Systems Track Lead
Spent eight years on model serving infrastructure at two regional technology companies. Leads the AI Systems Engineering track and architecture reviews.
Krishnan Lim
Programme Coordinator
Manages cohort scheduling, learner communications and track logistics. Previously coordinated technical training programmes at a Cyberjaya-based engineering firm.
Nurul Izzah
Mentor · Deep Learning Track
Computer vision engineer who works on production image processing systems. Mentors learners through the Applied Deep Learning track's project milestones.
Farhan Ooi
Mentor · ML Engineering
Data scientist with experience in evaluation frameworks and model monitoring. Provides written feedback on Foundations assignments and advises on project design.
Ruhani Chong
Content & Learning Design
Builds the written materials, recorded sessions and assessment frameworks for each track. Background in technical writing and instructional design.
How We Keep Quality Consistent
Code Review on Every Assignment
All graded work receives written feedback from a practising engineer, not an automated system. Comments address both correctness and approach.
Cohort Sizes Kept Small
Study groups are capped at eight to twelve learners so that mentors can give meaningful attention to each person's questions and project work.
Curriculum Reviewed Each Term
Track content is reviewed before each new cohort. Sections that become dated — libraries, APIs, best practices — are updated before the next term begins.
Data Privacy Practices
Learner data is collected only for programme administration. We do not sell or share personal information with third parties. See our Privacy Policy for full detail.
Milestone-Based Assessment
Progress is assessed through project milestones and assignments rather than timed examinations, which gives a more accurate picture of practical capability.
Post-Track Feedback Collected
Each cohort is surveyed at completion. Responses are shared with the curriculum team and inform adjustments to the next version of the track.
What Guides the Teaching Work
AI engineering education in Malaysia has expanded considerably over the past few years, but most of what's available sits in one of two categories: short workshops that introduce a concept without building it properly, or lengthy academic programmes that cover theory well but rarely reach the production concerns that practitioners face daily.
Tensorloom occupies the space between those two. The tracks are long enough to develop real capability — eight, sixteen and twenty-four weeks respectively — and they are structured so that each week's work connects to the weeks that follow it. A learner who completes the Foundations track will have written code that is reviewed by an engineer, built a project on a real dataset, and developed an understanding of evaluation that goes beyond accuracy scores.
The Applied Deep Learning track extends this into neural network territory, with PyTorch as the primary framework and an emphasis on the parts of the work that take longest to learn from self-study: understanding why a training run is behaving the way it is, constructing a dataset that will produce a useful model, and diagnosing errors systematically rather than by trial and error.
The AI Systems Engineering track addresses the concerns that arise after a model works in a notebook and needs to work in a product: latency, throughput, retrieval, monitoring, and the accumulation of small decisions that determine whether a deployed system stays useful over time. This track includes a capstone project built against a real brief and defended before a small panel.
All tracks are delivered online, in the evening, in Malaysian time. The compute-intensive tracks include GPU and cloud credits so that learners are working on the same infrastructure they would encounter at work, not on constrained hardware that changes how the problems look.
Curious About a Track?
Send an enquiry and we'll answer questions about prerequisites, schedule and what each track covers week by week.
Get in Touch