Engineering education built for the work, not around it
Latent Bazaar is an online school based in George Town, Penang. We run structured AI development courses for engineers who need to build working systems — and know what they built.
← Back to HomeOur Story
Latent Bazaar started from a specific frustration: engineers in Malaysia who wanted to develop AI skills could find plenty of material, but almost none of it was calibrated to the kind of industrial and enterprise engineering work that actually needed doing here. The courses that existed either aimed at researchers or bundled up beginner Python instruction as if technical readiness did not matter.
We set out to build something different. The name borrows the image of a covered market — a place where each stall is self-contained, its wares plainly labelled, its scope not inflated to fill a hall. A learner walks in, reads the sign, checks whether the stall suits them, and either steps inside or moves on. Nobody stands at the entrance promising the stall solves everything.
The first two courses — Practical Computer Vision Engineering and Retrieval Systems and Model Evaluation — were written for engineers with hands-on Python experience who are already working near the problems these courses address. The Enterprise AI Enablement Programme came later, in response to requests from Malaysian technical teams who wanted a structured way to build capability on an internal problem rather than a generic curriculum.
We are based in George Town, Pulau Pinang, and our office is open Monday through Saturday. The courses themselves run online with live sessions and weekly office hours to keep the learning tethered to real engineering problems rather than drifting into abstraction.
What We Teach
Engineering practice for AI systems. We teach how to build, measure and maintain vision models, retrieval pipelines and evaluation harnesses — not how to use a no-code dashboard.
Who Enrols
Python engineers, data engineers, and technical teams at Malaysian companies. The entry bar is stated on each course sign. We do not accept enrolments from people the course would not serve.
How We Work
Each course is a fixed scope with stated prerequisites, weekly hour ranges, a clear delivery record, and explicit limits on what the course does not cover. No implied promises.
The People Behind the Stalls
Each team member teaches the material they have spent time working with in actual engineering environments.
Syarifah Lim
Seven years writing and reviewing vision pipelines for manufacturing quality-control lines in Penang and Johor. Built the Computer Vision Engineering curriculum from first drafts and runs the weekly live sessions.
Arjun Rajan
Previously built search and recommendation infrastructure at two Malaysian fintech companies. Wrote the adversarial evaluation exercises that are central to the Retrieval Systems course.
Norhayati Zulkifli
Coordinates the scoping study and tailored curriculum for enterprise cohorts. Former engineering manager at a Malaysian bank; understands procurement, data governance and what teams in regulated industries can actually deploy.
Standards We Hold Ourselves To
These are the practices we apply across every course we run — not aspirations but operating habits.
Code Review on Every Submission
All assignment submissions receive a written review from an instructor, not automated scoring. Feedback is specific to the code submitted, not generic comments.
Scope in Writing Before Enrolment
What a course covers and what it does not are both written down before anyone pays. We do not narrow scope after fees are collected.
Malaysian Data Governance Alignment
The Enterprise Programme includes a data governance module aligned with published Malaysian requirements — reviewed each intake to reflect current obligations.
Learner Data Handled With Care
Submission data and assessment records are stored for the duration of the course plus 12 months, then deleted. We do not sell learner data or share it with third parties.
Written Assessment Records
Completers receive a written assessment record that describes what they did, not a generic certificate. It is a more honest document and a more useful one in a technical hiring conversation.
Feedback Collected and Acted On
Each cohort fills in a structured mid-course and end-of-course review. Changes to materials that result from that feedback are noted in the course changelog.
AI Development Training for Engineers in Malaysia
Malaysia's technology sector has expanded faster than the available pool of engineers who can build and evaluate AI systems from first principles. Vendor-led training covers specific tools. University electives cover theory. What is frequently missing is the middle ground: structured, hands-on course work that teaches someone who already writes code how to build, measure and maintain a specific class of AI system.
Latent Bazaar occupies that middle ground. Our three courses are calibrated to engineers at different stages: someone entering computer vision work for the first time but already competent in Python; someone who has deployed an AI feature but cannot yet evaluate it honestly; and a team of engineers who need to build AI capability on a specific internal problem under Malaysian data governance constraints.
George Town is a practical base for this work. Penang's manufacturing corridor generates real demand for vision engineering. The financial and technology companies concentrated in Kuala Lumpur and the Klang Valley provide the enterprise context for retrieval and evaluation work. We are aware of both environments when writing course materials and scoping enterprise programmes.
We do not offer placement, salary projections or career outcomes. We teach what the course says it teaches, to people who already have the foundation to use it. That is a narrower promise than many online schools make, but it is the one we can keep.
Ready to look at the stalls?
Browse the three courses or send us a message and we will point you toward the one that fits where you are technically.