Todak Academy · Studios · Fusion ← Home  ·  Documentation →  ·  AI Creator Program · Game Track
Module Specification · Draft v1

AI-Assisted Game Development

A 12-week, competency-based module that takes a beginner from “what is an AI agent?” to a real game published on the store — built on the way Todak actually works: Claude Code paired with a memory that never forgets the project.

Duration
12 weeks3 months · blended
Training hours
~240 hrsindicative, guided
Format
CoCUJPK competency units
Graduate ships
1 gamelive on store
Target qualification
Micro-credential cluster → SKM Level 3Game / Interactive Multimedia · accreditation-ready, pending JPK
Assessment
CompetencyCompetent / Not Yet
Entry
Openno coding needed
01

The method is the spine

This is what separates the module from any other bootcamp. It runs through every competency unit below.

Claude Code + a persistent “project brain.”

Each learner works with Claude Code as their AI pair, backed by their own pgVector project memory — a store that holds the game design, the code conventions, and every decision and bug they hit. The AI never loses the thread of the project, which is the only reason a beginner can ship in three months. Learners don’t just prompt — they run the loop, every session:

01
Prompt
→
02
Plan
→
03
Build
→
04
Verify
→
05
Save to memory
Not prompt-only. Not vibe coders.

AI writes a lot of the code in this course, and learners are taught to direct it well. That is not the finish line. A graduate can open their own game with the AI switched off, read the code, explain why it works, fix it when it breaks, and ship it to a store. The AI is the pair, not the pilot. Prompting is a skill we teach and grade. Understanding what comes back is the skill that makes it worth anything.

02

Competency profile

Eight competency units in JPK CoCU structure. Core units are required; the 3D pathway is elective, so a learner can ship a 2D game and still complete the module.

CodeCompetency unitTypeWeeksHours
AGD·C01AI-Assisted Development FoundationCore124
AGD·C02Game Design FundamentalsCore220
AGD·C03Game Programming with GodotCore3–440
AGD·C04Game Application Development & BuildCore5–636
AGD·E053D Asset Creation PipelineElective · 3D728
AGD·E06Character Rigging & AnimationElective · 3D8–932
AGD·C07Game Production — Vertical SliceCore10–1140
AGD·C08Game Publishing & DistributionCore1220
Total indicative training hours240
03

Delivery plan — week by week

How the units are taught across the term. Colour shows who leads each week.

Todak method — AI + self-driven Studios crew (Fareez & team) Mentor-led production
PHASE 1

Think in AI, design in games

Weeks 1–3 · Academy
01WK

How AI actually works — our way of using it

AGD·C01
Plain-language LLM / agent / agentic concepts, hands-on. Set up Claude Code + the pgVector project brain and run the first loop.
Working AI setup + initialised project brain
02WK

What makes a game fun

AGD·C02
Core loop, feedback, reward; genres and why scope matters. Study real games, pick a concept, write a 1-page design doc.
1-page GDD, saved to the project brain
03WK

Prompting with memory + first Godot steps

AGD·C03
Prompt Claude Code against the project brain; learn Godot nodes & scenes; build a Pong clone the Todak way.
Playable Pong clone
PHASE 2

Build in Godot — prototype to app

Weeks 4–6 · Fusion
04WK

Game logic, the AI-pair way

AGD·C03
GDScript essentials — signals, input, physics, collision — learned by building. Debugging as a real skill.
2D platformer — move, jump, die, win
05WK

From prototype to app

AGD·C04
Structure the project like real software: state machine, scene management, save/load, settings, menus, and git.
Multi-level game with full menu + save
06WK

First real build — ship to a phone

AGD·C04
The export pipeline to Android, testing on a real device, fixing with memory of past bugs.
Your game running on your own phone
PHASE 3

Make it 3D — creation & rigging

Weeks 7–9 · Fusion · elective
07WK

3D creation & the AI asset pipeline

AGD·E05
AI image → 3D model, Blender cleanup, style consistency, import to Godot. Runs on Todak’s own GPU.
First AI-made 3D asset, cleaned & in-engine
08WK

Rigging & animation

AGD·E06
Armatures & skeletons, weight-painting basics, a character rig, animation states, clean export to Godot.
A rigged character animating in-game
09WK

Assets into gameplay

AGD·E06
Wire the character into a real scene: animation state machine, camera, basic 3D movement. Decide 2D vs simple-3D for the ship.
Playable micro-scene with your character
PHASE 4

Slice & ship

Weeks 10–12 · Studios
10WK

Vertical slice

AGD·C07
Lock scope to one core loop, build it end-to-end, add game feel (juice, particles, transitions) and audio.
Vertical slice — one polished loop, start to end
11WK

Polish & pre-publish

AGD·C07
Bug bash, optimization, and the store listing kit — icon, screenshots, trailer, copy — plus AI disclosure and age rating.
Release candidate + full listing kit
12WK

Publish — through Todak Studios

AGD·C08
Walk the publishing runbook, submit through the Todak Studios developer account, pass the quality-gate check.
🚀 A published game with the learner’s name on it
04

Competency units in detail

Each unit follows CoCU structure: work activities, related knowledge and skills, and the evidence a learner must produce to be judged Competent.

AGD·C01

AI-Assisted Development Foundation

24 hrs · Core
Operate the Todak AI method — Claude Code with a persistent project memory — to plan, build, verify and document work.
Work activities
  • Set up the AI development environment (Claude Code + pgVector memory)
  • Apply the prompt → plan → build → verify → save loop
  • Maintain project memory: design, conventions, decisions, defects
Related knowledge & skills
  • What an LLM, agent and agentic system is; context & memory
  • Structuring a prompt; recalling and verifying against memory
  • Version-control basics
Evidence: a working environment and one memory-driven task completed and documented in the project brain.
AGD·C02

Game Design Fundamentals

20 hrs · Core
Analyse game types and produce a scoped, buildable game design document.
Work activities
  • Analyse genres, core loops and player motivation
  • Evaluate scope and feasibility for a 12-week build
  • Define a concept and produce a 1-page GDD
Related knowledge & skills
  • Core loop, feedback and reward; genre conventions
  • Scope discipline — why small games win
  • Design-document writing
Evidence: an approved 1-page game design document.
AGD·C03

Game Programming with Godot

40 hrs · Core
Build 2D game logic in Godot 4 with GDScript, AI-assisted throughout.
Work activities
  • Use the Godot editor: nodes and the scene system
  • Write GDScript: variables, signals, input, physics, collision
  • Implement sprites and animation; debug systematically
Related knowledge & skills
  • Programming logic fundamentals
  • Godot architecture; event-driven signals
  • Reading errors and verifying AI-assisted fixes
Evidence: a playable 2D platformer with lose and win states.
AGD·C04

Game Application Development & Build

36 hrs · Core
Structure a game as a real application and produce a build that runs on a device.
Work activities
  • Implement state machine, scene management, save/load, settings, UI/HUD
  • Apply version control to the project
  • Export to Android and test on a physical device
Related knowledge & skills
  • Application architecture and game state
  • Persistence and configuration
  • Build & export pipeline; on-device debugging
Evidence: a multi-level game with menu and save, running from an installable build.
AGD·E05

3D Asset Creation Pipeline

28 hrs · Elective (3D)
Produce game-ready 3D assets using an AI-assisted pipeline on Todak’s own hardware.
Work activities
  • Generate a 3D model from AI and evaluate output
  • Clean up and retopologise in Blender
  • Maintain art-style consistency; import into Godot
Related knowledge & skills
  • 3D asset fundamentals; mesh & texture
  • AI generation workflow and its limits
  • Blender cleanup; engine import settings
Evidence: one AI-generated 3D asset, cleaned and running in-engine.
AGD·E06

Character Rigging & Animation

32 hrs · Elective (3D)
Rig and animate a 3D character and integrate it into the game engine.
Work activities
  • Build an armature/skeleton and apply weight-painting
  • Create animation states — idle, walk, jump
  • Export the rigged character cleanly to Godot
Related knowledge & skills
  • Rigging concepts; bones and weights
  • Animation state machines
  • Skeletal export and engine integration
Evidence: a rigged character animating inside a Godot scene.
AGD·C07

Game Production — Vertical Slice

40 hrs · Core
Produce a polished, end-to-end vertical slice ready for review.
Work activities
  • Lock scope to one core loop and build it fully
  • Apply game feel: juice, particles, transitions
  • Integrate audio and SFX; iterate via weekly builds
Related knowledge & skills
  • Production discipline and scope control
  • Game feel and polish techniques
  • Playtesting and iteration
Evidence: a polished vertical slice / release candidate.
AGD·C08

Game Publishing & Distribution

20 hrs · Core
Prepare and publish a game to a distribution platform in compliance with platform and AI-disclosure policy.
Work activities
  • Prepare the store listing kit — icon, screenshots, trailer, copy
  • Complete AI-content disclosure and age rating
  • Submit via the publisher account and verify the game is live
Related knowledge & skills
  • Store policies and submission flow
  • AI-disclosure and compliance requirements
  • Release readiness and quality-gate criteria
Evidence: a game published (or quality-gate-ready) with a complete store listing.
05

Tools & rough cost

The stack is chosen to keep learner cost near zero. The engine, 3D tools, rigging, editing and version control are all free — the only real spend is one or two AI subscriptions.

Free forever · RM0 each
Godot 4 game engine · no royalties RM0 Blender 3D + rigging RM0 Mixamo auto-rig + animations RM0 DaVinci Resolve trailer / video RM0 Audacity audio RM0 Git / GitHub version control RM0 Supabase the pgVector project brain · free tier RM0
The only real spend · AI subscriptions
ToolWhat it’s forPlan~USD / mo~RM / moWhen
Claude (Claude Code)The core AI pair + memory workflowPro$20~95All 12 weeks
Meshy.aiAI 3D model generationFree → Pro$0–200–953D elective only (~4 wks)
2D image genSprites, textures, store artAcademy pool / Todak GPUshared—As needed
pgVector memoryThe “project brain”Supabase free tier$00All 12 weeks
2D track~RM285
with 3D pathway~RM380
Total AI + software cost per learner, for the whole 3 months. The 2D baseline is just Claude Pro for three months; the 3D pathway adds about one month of Meshy. Bought as Academy team/pool plans, the per-head cost drops below this.
Per cohort · 15–20 learners (pooled)
Line item15 learners20 learners
Claude Pro seats — core, all 3 monthsRM4,275RM5,700
Meshy.ai 3D — elective, pooled (~1 month)~RM800~RM1,000
Free tools — engine, rigging, editing, memoryRM0RM0
Publishing — on Todak Studios accountsRM0RM0
Cohort total (with 3D pathway)~RM5,100~RM6,700
effective / learner~RM335
A 2D-only cohort drops to ~RM4,275 (15) / ~RM5,700 (20) — no Meshy at all. Claude is roughly 85% of the bill and is basically fixed per active learner, so pooling mainly saves the 3D credits — and Todak’s own Higgsfield / Tripo pipeline on the in-house GPU can cover that 3D step, trimming the Meshy line further.

Store & publishing fees (Google Play, Steam, Apple, itch.io) are on Todak Studios’ developer accounts — a learner pays RM0 to publish. Figures are indicative 2026 retail prices (billed in USD; RM shown at roughly 4.75). Heavy Claude Code users may need the Max plan (~$100/mo), which is exactly why the pooled-account model exists.

06

Assessment & certification

Competency-based, evidenced by a real game

Every unit is judged Competent / Not Yet Competent against the evidence above — there are no written exams. The learner’s portfolio (their project memory, builds and assets) is the running record, and the published game is the capstone evidence, gated by the Todak Studios quality review.

Learners who complete all Core units earn the module credential; the 3D pathway (Elective units E05–E06) is stacked on top for those who take the 3D route. Units are designed to be individually recognisable as stackable micro-credentials.

07

TVET alignment

Structured to JPK CoCU — accreditation-ready

This module is written in the Curriculum of Competency Unit (CoCU) structure used by Malaysia’s Department of Skills Development (JPK/DSD): competency units split into work activities, related knowledge and skills, indicative training hours, and competency-based assessment. It is positioned as a stackable micro-credential cluster toward Sijil Kemahiran Malaysia (SKM) Level 3 in a Game / Interactive Multimedia occupational area, leveraging Todak Academy’s existing JPK accredited-centre status.

Status: this is a draft specification. Unit codes, hour allocations and the SKM mapping are indicative and become official only on JPK accreditation. It is not yet an accredited NOSS, and the accreditation pathway (and any MARA / PTPK / HRD Corp funding) is a separate step.