Latentgrove
Enquire Now
AI course solutions

Three Tracks

The full path from first principles to a deployed system

Each course is complete on its own. Together they form a continuous path — roots underground, branches reaching out, a canopy where the work becomes visible.

← Back to Home

Our Approach

How the three tracks work as a system

The three courses are not three separate products you can mix and match freely. They are stages in a designed sequence, each one requiring a different level of preparation and delivering a different kind of outcome.

You can enter at any stage if you can demonstrate the prerequisite knowledge. You cannot skip a stage and expect the next one to make sense — we will tell you this clearly during the enquiry process rather than after you have enrolled.

The sequence ends with a real system that a learner built, under their own name, with a practising engineer reviewing their architectural decisions. That is the outcome the whole structure is designed to produce.

Stage 1: Roots

Python and statistics. The two foundations everything else stands on. Seven weeks, six hours per week.

Stage 2: Branches

Applied deep learning. Neural networks, computer vision, transformers, and the engineering that makes experiments reproducible. Fifteen weeks, eleven hours per week.

Stage 3: Canopy

Build and ship one system with a practising engineer as mentor. Twenty-two weeks, ending in a technical report and a hosted project page.

Roots: Programming and Statistics
Track 1 · 7 weeks · 6 hrs/week

Roots: Programming and Statistics

RM 465

A seven-week course laying the two roots every later track depends on: Python written to be read by others, and the statistics required to say whether a result means anything. Written for beginners, including those returning to study after many years. Six hours a week, recorded and live.

What is covered

  • Python syntax, data structures, and readable code style
  • Probability, distributions, and statistical significance
  • Numpy and Pandas for data handling
  • Exploratory analysis and visualisation
  • Version control basics and reproducible notebooks

Process

1

One recorded lecture and one live session per week covering new material

2

Weekly exercise submitted through the forum, returned with written notes within four working days

3

Two office-hour sessions per week where you can bring unresolved questions

4

Written record of course completion issued after the final week

Enquire About Roots
Track 2 · 15 weeks · 11 hrs/week

Branches: Applied Deep Learning

RM 1,880

A fifteen-week track through neural networks, computer vision, sequence modelling, transformers, fine-tuning and evaluation, taught alongside the engineering practices that keep experiments reproducible. Suited to learners who have completed a foundations course. Eleven hours a week with two evening sessions.

What is covered

  • Feedforward networks, backpropagation, and training dynamics
  • Convolutional networks and computer vision tasks
  • Recurrent networks and sequence modelling
  • Transformers, attention, and fine-tuning pre-trained models
  • Evaluation methods and reproducible experiment tracking

Process

1

Two live evening sessions each week covering lecture and workshop material

2

Four assessed projects submitted at intervals — each reviewed with tutor code notes

3

GPU credits allocated at the start of the track for training on real hardware

4

Written record of course completion and full project submission history issued at close

Enquire About Branches
Branches: Applied Deep Learning
Canopy: Capstone and Mentorship
Track 3 · 22 weeks · with mentor

Canopy: Capstone and Mentorship

RM 4,070

A twenty-two week programme in which each learner builds and ships one system with a practising engineer as mentor, then writes and defends a technical report. Fortnightly architecture review, monthly cohort presentation, and workshops on writing, portfolio presentation and interview practice.

What is covered

  • Problem scoping and architecture decision-making with mentor
  • Data pipeline design, training runs on cloud hardware
  • System deployment and basic inference infrastructure
  • Technical writing and report structure
  • Portfolio presentation and interview preparation workshops

Process

1

Weeks 1–4: Project scoping with mentor, initial architecture proposal reviewed fortnightly

2

Weeks 5–18: Build and train the system; fortnightly check-ins with mentor; monthly cohort presentations

3

Weeks 19–21: Technical report drafted, reviewed and revised

4

Week 22: Report defence; hosted project page published; completion record and alumni access granted

Enquire About Canopy

Choosing Your Track

Which course fits where you are

Feature Roots Branches Canopy
Duration 7 weeks 15 weeks 22 weeks
Weekly hours 6 hrs 11 hrs varies
Prior knowledge needed None Python + statistics Deep learning fluency
Assessment type Weekly exercises 4 graded projects Built system + report
GPU / cloud compute —
Practising engineer mentor — —
Hosted project page — —
Fee (RM) 465 1,880 4,070

Best for

Roots

Someone with no programming background, or someone who has some exposure to Python but no statistics, who wants to prepare properly before touching any machine learning material.

Best for

Branches

Someone who can write Python comfortably and has a working understanding of probability and linear algebra, and who wants to move into deep learning with real assessed projects and code review.

Best for

Canopy

Someone who has completed Branches (or equivalent) and is ready to build and ship one real system, produce a technical report, and have a concrete, demonstrable project in their portfolio.

Standards

Shared across all three tracks

Data and privacy

Learner data is held on Malaysian infrastructure and is not shared with third parties outside of legal obligations. Submission records are kept for two years and then deleted.

Reproducible work

All assessed work must specify the environment, dataset version and random seeds. Reproducibility is a skill we teach and a standard we apply to every submission.

Named tutors for each cohort

You know who is reviewing your exercises and can address follow-up questions to that person directly. Feedback is not assigned to whoever is available — it is the same reviewer throughout your track.

Curriculum updated each cohort

After each cohort closes, the teaching team reviews what needed extra time and what the field has shifted. Changes are applied before the next intake. We do not run the same material unchanged for years.

Honest completion records

Completion records describe what was taught, how it was assessed, and the learner's standing. They do not claim academic qualification status. We say clearly what they are and are not.

Cohort size enforced

Maximum eighteen learners per cohort. This is not a soft guideline — intakes close when that number is reached. We do not increase cohort sizes to meet demand.

Fees

Complete pricing — no add-ons

All included resources (compute, forum, completion record) are part of the stated fee.

Roots

RM 465

7-week course

  • Recorded + live sessions (6 hrs/week)
  • Weekly exercises with written feedback
  • Twice-weekly office hours
  • Staffed cohort forum
  • Written completion record
Enquire
Branches
Most detailed

RM 1,880

15-week course

  • Two live evening sessions per week
  • Four assessed projects with code review
  • GPU compute credits included
  • Cohort forum
  • Written completion record
Enquire
Canopy

RM 4,070

22-week programme

  • Dedicated practising engineer mentor
  • Cloud credits for training and deployment
  • Hosted project page published
  • Alumni forum access after completion
  • Written completion record
Enquire

Not sure which track fits where you are?

Send an enquiry and describe your background. We'll come back with a direct recommendation — no pressure in either direction.

Send an Enquiry