Tensorloom
Stacked tensor layer diagram

Course catalogue

Three courses, each with a defined scope, stated prerequisites and a full syllabus

Python Foundations for ML · RAG Workshop · Applied Deep Learning Track. Full details on what each course covers, what it does not, and what you need to know first.

Back to Home

How the courses are structured

Teaching methodology in plain terms

Prerequisites first, syllabus second

Every course page opens with a plain description of what you need to know before starting. A self-check quiz helps you decide whether the course is at the right level.

Workload stated as a segmented bar

Hours per week are shown as a static figure on the page — not as a vague range. You can plan your schedule before committing to a cohort.

"What this does not cover" is written out

Each course has a section listing topics that are deliberately out of scope. This is placed above pricing on the course page so you see it before you decide to enquire.

Course 01 · 8 weeks

Python Foundations for Machine Learning

Prerequisites

Comfortable with Python loops, functions and dictionaries. A short self-check quiz is available to help you decide honestly before enrolling. No machine learning background required.

An eight-week evening course covering the specific subset of Python that machine learning work actually uses. NumPy array thinking and vectorised operations, pandas for tabular data, matplotlib for visual inspection, and the habits — virtual environments, version control, notebooks that still run tomorrow — that keep later work from collapsing.

Taught live in small cohorts, recorded the same evening. Scope is deliberate: this course does not cover machine learning algorithms or model training. It covers the Python layer underneath those things.

What this course does not cover

  • — Machine learning algorithms or model training
  • — Deep learning or neural networks
  • — Deployment or production infrastructure
  • — Data engineering pipelines at scale

What is included

  • Live sessions (2 evenings per week × 2 hours) + same-evening recordings
  • Weekly exercise sets with written feedback from the instructor
  • 3 portfolio-quality notebooks you keep outright
  • Cohort discussion channel for the duration of the course
  • Roughly 4 hours of homework weekly
Python and NumPy array operations

Week-by-week structure

01

Environment setup and Python data model

Virtual environments, pip, version control basics, Python's object model as it applies to arrays.

02

NumPy: arrays, axes and broadcasting

Array creation, reshaping, broadcasting rules, indexing patterns used in ML code.

03–04

Vectorised operations and performance

Writing operations without Python loops, profiling, understanding where vectorisation matters.

05–06

pandas for tabular data

DataFrames, indexing, groupby, merging, common data cleaning patterns.

07

matplotlib and exploratory visualisation

Reading your data visually before modelling — distributions, correlations, anomalies.

08

Portfolio notebook review

Presenting and reviewing the three portfolio notebooks. Feedback on code quality and structure.

Vector search and chunk retrieval diagram

Workshop schedule (7 hours with breaks)

9am

Chunking strategies

Why chunking decisions matter more than most people expect — fixed-size, semantic, and recursive approaches.

10am

Embedding models

Choosing and comparing embedding models, understanding the trade-offs.

11am

Vector stores and hybrid search

Setting up a vector store, running hybrid search, understanding recall-precision trade-offs.

1pm

Reranking and evaluation

Adding a reranker, measuring whether answers are actually grounded, honest evaluation approaches.

3pm

Build session

You build one complete working system on your own machine. Instructor available throughout.

Course 02 · 1 day

Single-Topic Workshop: Retrieval-Augmented Generation

Prerequisites

Comfortable with Python and the command line. No machine learning background required. You should be able to run a Python script, install packages, and navigate a terminal.

A one-day Saturday workshop, live online, on building a retrieval system around a language model. The scope is deliberately narrow: chunking strategies, embedding models, vector stores, hybrid search, reranking, and evaluation of whether answers are actually grounded. You build one working system during the day on your own machine and leave with the repository.

What this workshop does not cover

  • — Model training or fine-tuning
  • — Deployment at production scale
  • — Prompt engineering as a general discipline
  • — Agent frameworks or tool use

What is included

  • 7 hours live workshop with breaks (Saturday, online)
  • Maximum 20 participants
  • Starter repository you keep
  • Evaluation notebook
  • Written reading list of primary sources
  • Session recording

Course 03 · 24 weeks

Applied Deep Learning Track

Prerequisites

Solid Python, comfortable with the command line, and either the Tensorloom Python Foundations course or equivalent hands-on experience. This track moves quickly — the Python layer is assumed, not taught.

The school's longest programme: a six-month, part-time track for engineers moving into deep learning work. Covers the mathematics you actually need rather than a full degree's worth, then works through training loops written from scratch, convolutional and transformer architectures, transfer learning, experiment tracking, data pipelines, distributed training basics, quantisation, serving, and monitoring in production.

Two-thirds of the time is spent on projects rather than lectures. Each learner ships four graded projects, each reviewed line by line by a practising engineer. The track closes with a capstone the learner defines themselves.

What this track does not cover

  • — A full degree's worth of mathematics
  • — Reinforcement learning
  • — Multimodal models beyond vision + text
  • — Hardware design or chip-level optimisation

What is included

  • Live sessions twice weekly + recordings
  • Four graded projects, each reviewed line by line by a working engineer
  • Capstone project defined by the learner, with written review
  • Mentor office hours each week
  • Portfolio repository you own outright
  • Roughly 12–15 hours per week commitment

Instalment arrangements may be available — mention it in your enquiry.

Neural network architecture diagram — layers and batches

Track structure overview

1–4

Mathematics and training loops

Linear algebra and calculus as needed, autograd, writing training loops from scratch without a framework abstraction.

5–10

Architectures

CNNs, attention mechanisms, transformer architectures, transfer learning and fine-tuning strategies.

11–15

Training infrastructure

Experiment tracking, data pipelines, distributed training basics, mixed precision.

16–20

Production

Quantisation, model serving, monitoring in production, handling distribution shift.

21–24

Capstone

Learner-defined project, built over the final four weeks and reviewed with the same written feedback as the graded projects.

Which course is right for you

Choosing between the three programmes

Feature Python Foundations RAG Workshop Deep Learning Track
Duration 8 weeks 1 day 24 weeks
Hours per week ~8 hrs 7 hrs (one day) 12–15 hrs
Live sessions
Engineer-reviewed projects exercise feedback only build session
Portfolio deliverables 3 notebooks 1 repository 4 projects + capstone
Price (RM) 1,480 620 4,700
Best suited to Engineers new to the ML Python stack Engineers wanting hands-on RAG in one day Engineers committing to deep learning work

Across all courses

Standards that apply to every programme

Learner data privacy

Name, email and payment details held for operational purposes only. Not shared with third parties for marketing.

Syllabus currency

Syllabi are reviewed before each intake and updated when tooling or standard practice has changed enough to make existing content inaccurate.

Code quality in materials

Example code is typed, version-controlled, and written to run six months from the session date — not just during the course.

Enquiry response time

Enquiries are answered within one business day, Monday to Friday. We are willing to answer scope and prerequisite questions before you decide to enrol.

Cohort cap enforcement

Caps are not increased when demand exceeds them. When a cohort fills, a waitlist opens for the next scheduled intake.

No upsells during a cohort

The price stated at enrolment is the full price. There are no additional materials, upgrade tiers, or supplementary purchases offered once a course has started.

Pricing

Course fees in Malaysian Ringgit

All prices are full fees — no extras added during enrolment. Payment by bank transfer or debit/credit card.

Python Foundations

RM 1,480

8 weeks · per enrolment

  • 16 live sessions
  • 8 exercise sets with written feedback
  • 3 portfolio notebooks
  • Cohort channel access
Enquire

Deep Learning Track

RM 4,700

24 weeks · per enrolment

  • 4 reviewed projects + capstone
  • Weekly mentor office hours
  • Portfolio repository
  • Instalment option available
Enquire

Not sure which course fits where you are right now?

Send us a message describing what you already know and what you are trying to do. We will reply with a plain recommendation.

Send an Enquiry