Built to Show You
Where You Stand in the Stack
Bayang Labs designs programmes around a single principle: learners work better when they can see the whole system and locate their current work inside it.
Back to HomeWhy Bayang Labs Exists
Bayang Labs was set up in Ara Damansara in 2022 by a small group of engineers and educators who had spent years watching the same problem repeat itself: people working with AI systems — on both the technical and non-technical side — did not have a shared map of what they were looking at.
Product teams would hand off requirements to engineers without a working model of what a language system could actually do under different conditions. Fine-tuning practitioners would ship adapted models without a principled view of when adaptation was the right move and when it was not. Learners finishing technical courses would arrive at interviews with impressive projects documented in ways that obscured the decisions rather than explaining them.
The programmes at Bayang Labs address each of these gaps directly. The introductory programme builds a working mental model for non-engineers. The fine-tuning programme builds practical adaptation skills for engineers who already know the basics. The portfolio programme helps both groups communicate what they have built and why.
The name comes from the Malay word for shadow or shade — something that sits behind the visible surface and defines its shape. The AI systems in use today have a great deal of structure that is not visible in the outputs alone. Understanding that structure is what these programmes are for.
Sessions are held in Petaling Jaya. The introductory programme includes reference material in both English and Malay, which reflects the working environment of most teams in the Klang Valley. Programme sizes are kept small enough that learners can ask questions that matter to their specific context.
Bayang Labs does not operate as a placement agency or employment service. The programmes teach and provide practice. What learners do with that preparation is their own path to take.
Who Runs the Programmes
Ahmad Nadzri
Programme Director
Designs the curriculum architecture and runs the fine-tuning sessions. Previously led applied ML work at a Kuala Lumpur fintech over six years.
Shirin Tan
Instruction Lead
Leads the introductory programme and develops the bilingual glossary material. Background in technical writing and product education for enterprise software teams.
Rajan Krishnan
Portfolio & Interview Coach
Runs the portfolio preparation sessions and mock interviews. Spent several years on hiring panels for technical roles before moving into coaching and education.
How We Work
Stated Assumptions
Every programme page opens with a single-sentence assumed-knowledge statement. Learners know what they need before they pay.
Written Feedback
Code reviews, project assessments and mock interview debrief notes are delivered in writing so learners have a record to reference later.
Data Privacy
Learner data — contact details, submissions, progress notes — is stored only as long as needed. No sharing with third parties for marketing.
Small Cohort Sizes
Programmes cap at a size that allows meaningful interaction. Mentor sessions are individual, not shared slots within a large group.
Honest Scope
Programmes describe what is taught and practised. No employment outcomes, no placement services and no credential claims beyond what we actually deliver.
Content Currency
The AI field moves quickly. Programme materials are reviewed before each cohort to reflect significant developments in tooling and method.
Structured Learning for a Layered Field
The AI development field has a visible surface — large language models, diffusion systems, retrieval pipelines — and a less visible internal structure of weights, training objectives, data decisions and deployment constraints. Most accessible courses teach the surface. Bayang Labs programmes teach the structure.
The layer-stack framing used in all Bayang Labs materials borrows from the way network engineers think about protocol stacks. Each layer has a defined role, takes specific inputs, and produces specific outputs. Understanding the interfaces between layers matters as much as understanding any single layer in isolation. When a language model behaves unexpectedly, the cause is often in the layer below the one being looked at.
This framing works for non-technical learners too. A product manager who understands that a language model is a probability function over tokens — not a search engine and not a person — will write better specifications and make better decisions about when to use these systems and when not to. The introductory programme builds exactly that understanding, without requiring any code to be written.
Bayang Labs operates from Ara Damansara in Petaling Jaya, within easy reach of the broader Klang Valley. The location is deliberate: the Klang Valley has a substantial and growing technical workforce, and the programmes are designed to be useful to people already working in that context, not people relocating to attend a course.
Ready to Locate Yourself in the Stack?
Send us a message and we will help you identify which programme fits where you are now.
Contact Bayang Labs