From First Principles to
Production Systems
Three sequential tracks covering the full arc of AI engineering work — each designed to stand alone, and each designed to prepare you for the next.
Back to HomeHow the Tracks Are Designed
Each track is built around a set of weekly threads — topics that recur across the weeks rather than appearing once and being left behind. A concept introduced in week two of the Foundations track shows up again in week five, applied to a harder dataset. The same concept appears in the deep learning track, where it takes a different form. By the systems track, you are using it to make decisions about production architecture.
The design principle is that understanding something once, under good conditions, is not the same as being able to use it reliably when things are messy. The woven structure of the curriculum is meant to address that gap.
Written materials for every module
Twice-weekly live sessions (MYT)
Graded assignments with feedback
Project milestones throughout
Foundations of Machine Learning
An eight-week evening track covering Python for data work, linear algebra and probability applied to models, classical supervised and unsupervised methods, feature preparation and honest evaluation. Written for developers and analysts with some programming background who want a solid base before neural networks. Live sessions run twice weekly in Malaysian time.
What's Included
- Recorded sessions available after each live class
- Six graded assignments with written feedback
- A small project on a public dataset
- Code review from a practising engineer
- Completion record listing topics covered
Week-by-Week Process
-
1
Python for data work — environment setup, NumPy and pandas patterns, loading and inspecting datasets
-
2
Linear algebra and probability — matrix operations, distributions and the intuitions used in model derivations
-
3
Supervised methods — regression, classification, regularisation and the decisions that shape a trained model
-
4
Feature preparation and evaluation — handling real data, building an evaluation framework, avoiding common pitfalls
Applied Deep Learning Engineering
What's Included
- GPU credits for coursework (included)
- Weekly live sessions
- Four project milestones reviewed individually
- A mentor assigned for the full term
- Study group of 8–12 learners
- Completion record with portfolio attached
What the Track Covers
-
1
PyTorch fundamentals — tensors, autograd, training loops and the debugging patterns that matter
-
2
Architectures — CNNs, sequence models, attention mechanisms and transformer design
-
3
Transfer learning and fine-tuning — when and how to adapt pretrained models to a new task
-
4
Distributed training and dataset work — multi-GPU patterns, dataset construction and systematic error analysis
Sixteen weeks building and training neural networks in PyTorch. Suited to learners who have completed Foundations or who already train models at work. The track moves from basic architectures through transformers and transfer learning to distributed training and dataset construction.
AI Systems Engineering Track
Twenty-four weeks on taking models into production. Covers serving architectures, batching and latency, retrieval pipelines, evaluation harnesses, monitoring, drift detection, cost modelling and the operational practices that keep a deployed system honest. Aimed at engineers who will own a system after it ships. Includes a capstone built against a real brief.
What's Included
- Cloud credits (included)
- Weekly mentor sessions
- Code review at every milestone
- Two architecture reviews with practising engineers
- Capstone defended before a panel
- Portfolio preparation sessions
- Detailed completion record describing capstone scope
What the Track Covers
-
1
Serving architecture — latency budgets, batching strategies, GPU inference and REST vs streaming patterns
-
2
Retrieval and evaluation — vector search, RAG patterns and building an evaluation harness that stays honest
-
3
Monitoring and drift — logging, alerting, data and concept drift detection and retraining pipelines
-
4
Cost modelling and capstone — infrastructure spend analysis, build-vs-buy decisions and a defended real-brief project
Which Track Fits You
Use this as a guide. If you're unsure, send an enquiry and we'll advise based on your background.
| Feature / Aspect | Foundations RM 690 |
Deep Learning RM 2,480 Most enrolled |
AI Systems RM 4,450 |
|---|---|---|---|
| Duration | 8 weeks | 16 weeks | 24 weeks |
| Weekly hours | ~10 hrs | 12–15 hrs | ~15 hrs |
| GPU / cloud credits | GPU | Cloud | |
| Named mentor | |||
| Study group | |||
| Capstone project | |||
| Best for… | Developers new to ML | Engineers training models | Engineers owning production systems |
Protocols That Apply to Every Track
Privacy & Data Handling
Learner submissions and personal data are stored securely and used only for programme administration. Not shared with third parties.
Assessment Integrity
All graded work is reviewed by a person, not an automated checker. Feedback is specific to the submission, not generic.
Curriculum Currency
Track content is reviewed before each new cohort. Libraries, APIs and practices that have moved on are updated before learners see them.
Enquiry Response
Enquiries are handled by the programme coordination team, not by automated systems. Response time is one working day.
Completion Records
Every track produces a written completion record describing what was studied, built and reviewed — issued at the end of the term.
Cohort Size Limits
Study groups are capped by design. When a cohort is full, the next intake date is offered rather than enlarging the group.
Track Fees in Malaysian Ringgit
Foundations of ML
8 weeks- Recorded sessions
- Six graded assignments
- Dataset project
- Code review
- Completion record
Applied Deep Learning
16 weeks- GPU credits bundled
- Four project milestones
- Named mentor
- Study group
- Portfolio completion record
AI Systems Engineering
24 weeks- Cloud credits bundled
- Two architecture reviews
- Capstone + panel defence
- Portfolio preparation
- Detailed completion record
Not Sure Which Track to Start With?
Send us a short note about your programming background and what you're trying to build. We'll suggest the track that fits.
Ask a Question