[1] Gradient Hall
Company Benefits Solutions Testimonials Get in Touch
Gradient Hall school environment

## about_gradient_hall

A school built around the work that actually takes time

Gradient Hall was set up to teach the parts of AI development that are under-represented in most courses: data preparation, honest evaluation, and building systems that run end to end.

← Back to Home
[2]

## school_story

How Gradient Hall started

Gradient Hall was founded in Kuala Lumpur by a small group of engineers who had each spent time in different roles — data engineering, model development, product — and kept noticing the same gaps. Most AI courses either assumed no coding background, or assumed deep academic ML training. Very few addressed the middle: someone who writes code every day and wants to work more thoughtfully with AI systems without switching careers or going back to university.

The school started with the data course because that is where most engineering hours disappear. It is uncommon to find a course that uses genuinely messy data rather than a cleaned teaching set. The evaluation seminar came from watching teams ship model updates without having any structured way to answer whether the new version was actually better. The development track emerged when it became clear that many engineers wanted to build complete AI systems — not just understand the concepts — and needed a structured path to do that part-time.

We are based at 61 Jalan Ampang and run all teaching online, which means the cohort is not limited to Kuala Lumpur. Participants come from across Malaysia and occasionally from other parts of Southeast Asia.

[3]

## mission_statement

# Gradient Hall exists to teach the parts of AI work that take the most time
# and receive the least structured instruction — data preparation, evaluation,
# and building systems that someone who did not build them can understand,
# run, and extend.

# We teach working engineers. We use real data. We review real code.
# We make no claims about employment or earnings.

[4]

## team

The people who teach and review

AR

Amirul Rashid

Programme Lead, Data & Infrastructure

Has spent eight years building data pipelines for production ML systems across financial services and logistics. Designed the data handling course syllabus and reviews all exercise submissions.

SW

Siew Wei Tan

Instructor, Evaluation & Systems

Works as a practising engineer on retrieval and agent systems. Leads the evaluation seminar and provides written project reviews on the development track.

NK

Nurul Kamarudin

Track Mentor, Capstone Reviews

Software engineer with a background in serving infrastructure and LLM deployment. Conducts fortnightly office hours for the development track and reviews capstone projects.

[5]

## teaching_standards

How we hold ourselves accountable

Real data in every exercise

No exercise uses a pre-cleaned teaching dataset. We source or construct genuinely messy files so that learners work through the same friction they will encounter in a real project.

Written, substantive feedback

Submissions receive line-by-line written feedback from a practising engineer, not a rubric score. Comments explain the reasoning behind suggestions.

Stated scope, no scope creep

Every programme lists what it does not cover as clearly as what it does. Changes to syllabus content between cohorts are communicated in advance.

Personal data handled carefully

Participant information is used only for course administration and enquiry responses. We do not share it with third parties for any other purpose.

Prerequisites stated honestly

We describe what you need to know before starting, not just what you will learn. The development track includes a brief technical conversation before enrolment to confirm fit.

Primary sources, not summaries

Reading lists point to original papers and documentation rather than secondary blog posts. Where a concept comes from an academic source, we say so.

[6]

## school_values_and_context

Teaching AI development in Malaysia

Malaysia has a substantial software engineering workforce and a growing number of teams building AI-adjacent products. What is less common is structured technical education aimed at engineers who are already working and want to deepen specific skills — not retrain from scratch.

Gradient Hall's three programmes cover distinct, practically useful areas: preparing data for AI workloads, evaluating whether model changes are improvements, and building AI systems end to end from foundations through deployment. Each programme is self-contained. There is no requirement to take them in sequence, and each has its own stated prerequisites.

All sessions are online and scheduled in the evenings to accommodate working schedules. The development track runs over six months at a pace designed to be manageable alongside a full-time role. We keep cohorts small so that office hours remain genuinely useful and the cohort channel stays signal-heavy.

[7]

## contact_prompt

Still have questions about the school?

Use the enquiry form and we will respond within two working days.

Send an Enquiry