Education Built Around Careful Observation
Nuvex was founded on a straightforward idea: developers learn performance reasoning better when they start by watching what a workload actually does, not by memorising specifications.
Back to HomeHow Nuvex Came About
Nuvex started in George Town, Penang, when a small group of engineers noticed a gap: most resources on GPU performance were either vendor-produced documentation aimed at specialised hardware teams, or generic cloud tutorials that glossed over the underlying mechanics.
What developers working on AI workloads in Malaysia actually needed was a way to build a mental model — a structured way of asking "why is this slow?" and reasoning from first principles about what the hardware is doing.
We built the programmes we wanted to have taken ourselves. The lab-notebook format came from the way good engineering notes work: observations first, interpretation second, conclusions third. That structure keeps learning grounded in what you can measure.
What We're Here to Do
Build reasoning skills, not rote knowledge
Performance tuning is a reasoning process. We teach the thinking pattern, so learners can apply it to workloads we haven't covered.
Keep materials you can actually reference
Every programme produces notes and guides that stay with learners after the programme ends. The point is long-term utility.
Serve engineering teams as a unit
Individual learning helps, but teams benefit most when they share vocabulary and mental models. Our coaching programme is designed with that in mind.
Who Builds the Programmes
Hafiz Zulkifli
Curriculum Lead
Spent six years profiling ML workloads for a Penang-based hardware firm before joining Nuvex to translate that work into structured learning materials.
Rania Noor
Lab Design & Facilitation
Background in systems engineering and technical education. Designs the sandbox exercises and facilitates team coaching sessions at Nuvex.
Jia Wei Tan
Research & Content
Focuses on keeping programme content current as tooling and hardware evolve. Researches workload patterns for new lab exercises.
How We Maintain Programme Quality
Every programme at Nuvex follows a set of principles we review regularly to make sure the material stays accurate and the learning stays effective.
Content Accuracy Reviews
Materials are reviewed when significant changes occur in tooling or hardware generations that affect how concepts apply in practice.
Data and Privacy Handling
Learner information is handled in accordance with Malaysia's Personal Data Protection Act 2010. We do not share data with third parties for commercial purposes.
Exercise Validation
Lab exercises are tested against the sandbox environment before each cohort to confirm that expected observations still hold.
Participant Feedback
We collect structured feedback after each programme. Observations that point to unclear explanations or outdated examples are addressed in subsequent revisions.
Clear Scope Boundaries
Each programme defines clearly what it covers and what it does not. We do not position any track as a complete preparation for topics outside its stated scope.
Transparent Engagement Terms
For team engagements, scope, duration, and what's included are agreed in writing before the programme begins. No surprises mid-engagement.
What Guides Our Work
Nuvex operates on the principle that engineering education works best when it starts from what is observable — not from what the documentation says should happen. AI workload performance is complex, and the gap between "theoretically optimal" and "what your workload actually does" is where most of the interesting work lives.
Being based in George Town, Penang puts us in a region where hardware manufacturing and software development intersect in interesting ways. Our programmes are shaped by that context: practical, grounded in how workloads behave in real infrastructure, and written to serve engineers who don't have time for content that isn't directly useful.
See the Programmes We Offer
Three structured paths — from individual fundamentals to team-scale coaching. Find the one that fits where you are now.