Serious AI Education.
Built for Practitioners.
We run focused, structured programmes for people who want to develop working knowledge of artificial intelligence — not credentials to display, but skills to put to use.
Back to HomeHow Cogniva Came Together
Cogniva started from a fairly specific frustration. A group of software engineers working in Kuala Lumpur had the programming background to build things, but kept running into the same wall: the mathematics in research papers was opaque, and most available courses were either too shallow to be useful or too academic to be applicable.
The first programme — what is now the Maths Refresher — grew out of internal study sessions held in Bangsar in early 2022. The format worked because it was built around notebooks and problems, not slides and lectures. People could see where the notation connected to code they had already written. That combination of concreteness and structure turned out to be what was missing.
Over the following two years, the model-building curriculum took shape through a longer process of iteration. The core insight was that most learners in the region had access to good tutorials but very little exposure to the kind of critical feedback that comes from having experienced people review actual work. The critique format in the Studio programme is a direct response to that gap.
The Portfolio and Practice Programme was added when several learners who had completed earlier modules asked for something more sustained — a place to work on projects that could stand up to professional scrutiny and that came with a written record of what had been done and how.
Cogniva is based at 6 Jalan Telawi in Bangsar. Most sessions run online, but the address reflects where the organisation is actually rooted. The team is small and the programmes are designed to stay that way — cohort sizes are kept manageable so that critique and mentoring are substantive rather than nominal.
Our Purpose
To give working professionals in Malaysia a clear, structured path from knowing how to code to knowing how to build, evaluate and document AI systems that hold up under review.
Specificity over breadth
Each programme covers a defined scope thoroughly rather than surveying a wide field superficially.
Artefacts over assessments
Learners leave with notebooks, decision logs and portfolio projects — not just a score on a test.
Honest feedback
Critique sessions are direct and open. The goal is to surface what isn't working, not to reassure.
People Behind the Programmes
Syahrul Razali
Programme Director
Built and maintained production ML systems in the logistics and fintech sectors for nine years. Designed the Maths Refresher curriculum and oversees the critique format in the Studio.
Nurul Chong
Curriculum Lead
Specialises in applied NLP and data engineering. Developed the technical modules in the Portfolio and Practice Programme, including the data pipeline and deployment components.
Azrul Fadzil
Mentor Coordinator
Manages the mentor pairing process and interview practice workshops. Previously worked as a technical hiring manager and brings that perspective directly into the portfolio assessment process.
How We Work
Cohort Size Limits
Each programme cohort is kept small enough that critique is substantive. We would rather run a second cohort than dilute the quality of feedback on the first.
Data Privacy
Learner data is used only for programme delivery and communication. We do not share personal information with third parties for marketing purposes.
Consistent Curriculum
Programme content is reviewed and updated before each new cohort. Learners in one cohort work to the same syllabus and standards as those who follow.
Structured Mentoring
Mentor relationships in the Studio and Portfolio programmes are built around specific checkpoints, not open office hours. Each meeting has a defined purpose.
Honest Outcome Statements
We describe what learners will practise and what artefacts they will produce. We do not make claims about employment or income that we cannot substantiate.
Cloud Resource Management
GPU and cloud credits are allocated per cohort and managed through accounts we set up at enrolment. Usage is transparent and linked to programme assignments.
Applied AI Education in Malaysia
Cogniva operates in the space between introductory programming courses and full-time academic programmes. Most people who come to us have some coding background and an interest in AI, but haven't found a structured way to move from tutorials to building things that work in a real context.
Our programmes draw on what practitioners actually need: the mathematics that underpins gradient-based training, the engineering thinking required to design evaluations that could genuinely fail, and the documentation habits that make work legible to others. These are skills that take time and practice to develop, not information that can be consumed in a weekend.
Being based in Kuala Lumpur matters to how we build the programmes. Session times are set for Malaysian working hours. The problems we use in the curriculum come from contexts familiar to professionals in the region — logistics, financial services, e-commerce, manufacturing. The language of the programmes is English, but the frame of reference is local.
Cogniva is deliberately small. We are not trying to serve tens of thousands of learners with automated content. We are trying to run a small number of well-designed programmes where the feedback is real and the work is taken seriously. That is the only model of education we think is worth building.
Get in Touch
If you'd like to know more about Cogniva's programmes or discuss which one fits your current level, reach out using the contact form on the main page.
Contact Us