Tensorhaus
Engineering workbench

// benefits

What Tensorhaus courses
offer that others do not

Practitioner content, open tooling, and a pace built around full-time employment — not the other way around.

Back to Home

// overview

Core Advantages

Built for Practitioners

Content is written for people who already hold a technical role. We do not begin with first principles of programming — we begin where a working engineer's knowledge ends and ML operations begins.

No Vendor Lock-in

All labs run on publicly available tools. Every dependency can be installed without a paid licence. What you learn applies to your own infrastructure, not only a specific cloud platform.

Weekly Format

One lecture and one lab per week. The pace is steady enough to sustain learning but does not demand you put other responsibilities on hold to keep up.

Reviewed Work

Assignments and reflections are reviewed by instructors and returned with specific comments. You are not submitting into a void; someone reads what you wrote and responds to it.

Priced for Malaysia

Course fees are set for the Malaysian market. The ten-week MLOps course costs RM 4,580; the five-week Tabular Data course costs RM 780. No currency conversion needed.

Reading as a Skill

The Friday session trains structured code reading — a habit that most curricula skip entirely. Understanding existing ML code is a daily part of engineering work; we treat it as a subject.

// expertise

Instructor Background

Tensorhaus instructors have worked in production ML and data engineering roles. They have encountered the failure modes of deployed models, debugged data pipeline issues at 2 am, and written the documentation that kept a system maintainable six months after initial deployment. The course content reflects that experience, not a sanitised version of it.

  • Six or more years of production ML or data engineering experience per instructor
  • Domain coverage across fintech, e-commerce, and logistics sectors in Southeast Asia
  • Course materials reviewed and revised based on cohort feedback each year

// methodology

How the Courses Are Structured

Each week of the MLOps and Tabular Data courses has three parts: a recorded lecture, a lab notebook in a containerised environment, and a short written reflection. The lecture introduces the topic; the lab applies it to a concrete problem; the reflection asks you to articulate what you understood and where you are uncertain. That uncertainty is useful — it is the input instructors use when writing feedback.

  • Recorded lectures — watch when your schedule allows
  • Containerised labs — identical environment for every participant
  • Written reflections reviewed and returned within one week

// tooling

Open-Source Throughout

The MLOps course uses tools from the public ML engineering ecosystem: experiment tracking with MLflow, containerisation with Docker, and lightweight deployment patterns you can reproduce on your own infrastructure. The Tabular Data course works in Python with scikit-learn, LightGBM, and pandas. No proprietary platform is required for either course.

  • MLflow, Docker, Python — no paid platform dependency
  • Labs run on a personal laptop with 8 GB RAM
  • All course software is maintained as open-source projects

// support

What Support Looks Like

Each course has an async discussion channel where participants can post questions. Instructors check it on weekday mornings and reply with technical answers, not platitudes. Lab issues that cannot be resolved in writing are addressed in a weekly office-hours slot during the course period.

  • Async channel monitored weekday mornings
  • Weekly office hours during the course period
  • Lab setup questions handled before the course begins

// outcomes

What You Leave With

Completing the MLOps course means you have built and packaged a model, set up experiment tracking, written a deployment script, and documented the system well enough that another engineer could maintain it. These are concrete deliverables. The Tabular Data course leaves you with notebook examples of gradient-boosted models, validated with proper cross-validation, on three different datasets. The Friday sessions, taken over a period of months, develop the habit of reading code you did not write — a skill with no shortcut.

// comparison

Tensorhaus vs Typical ML Course Providers

Feature Typical Providers Tensorhaus
Prerequisite transparency Often vague or omitted Stated specifically per course
Tooling dependency Often tied to a paid cloud platform Open-source only, runs locally
Work reviewed by instructor Auto-graded or peer-reviewed only Instructor reads and responds to each submission
Employment claims Common feature of marketing None — scope is honest
Pace suits full-time employment Often intensive sprint format Weekly cadence, asynchronous lectures
Pricing currency USD pricing with conversion costs Malaysian Ringgit, local payment methods
Code reading as a subject Rarely addressed Dedicated weekly session

// what sets us apart

Distinctive Features

01

The Friday Reading Session

A structured weekly session built entirely around reading and annotating public ML code. No curriculum in Malaysia offers this as a standalone product. The habit it builds is the same one that separates engineers who can navigate unfamiliar codebases from those who cannot.

02

Containerised Labs, Pinned Dependencies

Every lab environment is defined in a Dockerfile with pinned dependency versions. Participants spend time learning, not debugging package conflicts. We test the environment before each cohort starts.

03

Instructor-Written Feedback

Reflections and assignment submissions are read by the course instructor, not by an automated system or a teaching assistant following a rubric. Feedback references what you specifically wrote.

04

Annual Content Revision

Course materials are reviewed annually. Sections that reference deprecated tooling or outdated patterns are rewritten. You study current practice, not a snapshot from three years ago.

// milestones

Where We Stand

3+

Years of course delivery

180+

Practitioners enrolled

96%

Lab completion rate across cohorts

4.7

Average feedback rating out of 5

MDeC Technology Training Partner

Recognised by Malaysia Digital Economy Corporation as a qualified technology training provider, 2024.

HRDF Claimable Courses

MLOps Fundamentals and Tabular Data Methods are claimable under the Human Resources Development Fund for eligible Malaysian companies.

KL Data Engineering Meetup Partner

Ongoing collaboration with the Kuala Lumpur Data Engineering community, hosting sessions since April 2023.

// enrol

Study the operational side of ML

If the structure and content here fit what you are looking for, send us an enquiry. We will reply with course details and answer your questions before you sign up.

Send an Enquiry