Gradient Owl
Gradient Owl study environment

Subang Jaya, Selangor · est. 2021

A school built around the hours you actually have.

← Back to Home

Our story

How Gradient Owl came together

Gradient Owl grew out of a straightforward observation: most people who want to move into AI development are already employed. They do not need a twelve-month full-time programme. What they need is something that fits between 8 pm and 10 pm on a Tuesday, with tutor feedback that arrives before Saturday's session.

The school was set up in Subang Jaya in 2021 by a group of engineers and educators who had spent years watching talented people drop out of online courses — not because the content was wrong, but because the pace and format did not fit around a job and a family. They built the evening structure first, and wrote the curriculum around it.

Three tracks now cover the full journey from a first Python script to a deployed, mentor-reviewed capstone project. Each one is designed to be taken in sequence, though learners who already have a programming background regularly enter at the second track without issue.

The name came later. An owl works at night. A gradient is what a model descends to learn. Put them together and you have a reasonable description of what the school actually is: a place where people do serious technical work in the dark hours, step by step.

The physical space in Subang Jaya is used for cohort workshops, peer review sessions, and the fortnightly capstone check-ins. Most teaching is delivered online so learners across Selangor and the Klang Valley can attend without commuting after a long day. The office is staffed during evening hours when questions are most likely to arise.

Our mission is unchanged from the first intake: give working adults in Malaysia a structured, honest path into AI development, with feedback from people who build models themselves.

The people behind the courses

Teaching team

AH

Ahmad Hafizan

Lead Instructor, Model Building

Former data engineer at a Kuala Lumpur fintech firm, Ahmad now writes and delivers the Model Building track. His focus is helping learners build models they can explain, not just models that score well.

NR

Nurul Rashidah

Course Author, Programming Groundwork

Nurul spent four years teaching data analytics at a polytechnic before joining Gradient Owl. She wrote the Groundwork curriculum specifically for people who have never opened a terminal before.

DL

Daniel Lim

Capstone Programme Director

Daniel oversees the Night Cohort Capstone, matching each learner with a mentor and running the peer review fortnights. He brings ten years of software engineering experience across Selangor and Singapore.

SY

Siti Yusmazida

Teaching Assistant, Groundwork & Model Building

Siti staffs the cohort forum on Tuesday and Thursday evenings. She completed both the Groundwork and Model Building tracks herself in 2022 while working in logistics, which makes her an unusually accurate judge of where learners get stuck.

KP

Krishnamurthy Pillai

Infrastructure & Cloud Tooling

Krishna manages the cloud environment that learners use during the Model Building and Capstone tracks. He maintains the GPU credit allocation and ensures the hosted project pages remain available after each term ends.

WC

Wei Chen

Curriculum Adviser

Wei reviews the Model Building track each intake and advises on transformer architecture content as the field moves. He works in natural language processing during the day and reviews course material on Saturday mornings.

How we run our courses

Standards we hold ourselves to

Written feedback on every exercise

Every submission receives written notes from a tutor within 72 hours — not a score, not a rubric tick-box. We write what was handled well and what to revisit, so learners have something to act on before the next session.

Personal data handled carefully

Enrolment and course data is stored on servers within Malaysia. We do not share personal information with third parties for marketing. Our data handling follows applicable Malaysian personal data protection obligations.

Capped intake per cohort

Each intake is limited so the teaching team can engage meaningfully with individual learners. We do not open a second cohort until we have the teaching capacity to staff it properly.

Curriculum reviewed each intake

The Model Building and Capstone content is reviewed before each new intake. The field moves quickly; we update transformer architecture coverage and evaluation design sections when the practice on the ground has clearly shifted.

Honest records of completion

We issue written records of course completion that describe what the learner studied and what they produced. We do not issue certificates that imply professional licensing or accreditation we do not hold.

Registered Malaysian entity

Gradient Owl operates as a registered business in Malaysia. Our physical address in Subang Jaya is staffed during evening office hours and is available for learners who prefer to attend in person for workshops and capstone presentations.

About our approach

Evening-paced AI education in Malaysia

The AI development skills market in Malaysia has grown considerably since 2020, but most structured learning options still assume the learner can dedicate a full working day to study. Gradient Owl was built specifically for people who cannot — engineers, analysts, managers, and career changers who carry a full professional workload and want to add practical machine learning skills alongside it, not instead of it.

The three-track structure — Groundwork, Model Building, Capstone — follows the logical sequence of skill development in the field. A learner who completes all three tracks will have written Python from scratch, built and evaluated supervised and unsupervised models, worked with neural network architectures and transformer-based systems, deployed a project to a cloud environment, and defended their work in writing before their cohort. That sequence mirrors what a junior AI developer would be expected to demonstrate in a technical interview at a Malaysian technology firm.

The evening delivery model shapes everything: session length, exercise scope, tutor response windows, and the cap on cohort size. We do not believe a two-hour evening session should cover the same volume as a six-hour daytime workshop. Each session moves through one concept thoroughly rather than several concepts quickly. Exercises are sized for the time between two evening sessions, not for a weekend dedicated to study.

Learners who join from a range of industries — logistics, banking, healthcare administration, civil engineering — often bring domain knowledge that enriches the cohort discussions. The diversity of professional backgrounds within each intake is treated as a feature of the course design rather than a complication to manage.

Want to know which track fits your background?

Send us a note with your current experience level and what you are hoping to build. We will suggest the right starting point without pressure to enrol immediately.

Start the Conversation