Generative AI Course

For creatives who want authorship over AI, not just output.

A 12-week studio for designers, architects, visualizers, and artists building real command of image generation, 3D workflows, motion, automation, and AI-assisted web creation.

12 Weeks

36 Guided Hours

Overview

Who it is for, and what it unlocks.

Overview / 01

Audience

People who tried ComfyUI and felt lost or overwhelmed.
Total beginners to node-based systems who want a clear mental model.
Creatives moving beyond drag-and-drop AI tools into reusable pipelines.
Artists, makers, and automation-focused teams exploring reproducible image and video generation.
Architects, interior designers, landscape architects, graphic designers, 3D artists, product designers, freelancers, and studios.

Objectives

Transform sketches and 3D massing into high-fidelity renders.
Build reusable custom AI workflows for repeated design tasks.
Rapidly iterate across styles, materials, mood, and lighting.
Animate architectural spaces and create cinematic motion outputs.
From builder to expert mastering complex ComfyUI pipelines and automation systems.
Combine ComfyUI with Krita, plugins, and zero-code web or app development flows.

Weekly sessions in a structured progression.

36h

Total guided instruction across the course.

Final project pipeline connecting concepts to delivery.

Instructors

Learn from practitioners building real AI workflows.

Mostafa Mohamed

AI Specialist, 3D Visualizer

Mostafa Mohamed is an AI Workflow Specialist focused on building systems, design, and visual production workflows for companies, with full-time experience at Stylus. He brings 2 years of hands-on work with ComfyUI, open and closed generative AI tools, and has contributed to projects with DSC, Emaar Masr, and Any Design Studio. He was also one of the former founders of Any More, an AI lab.

Mariam Adel

AI Workflow Specialist, Design Architect

Mariam Adel is an AI Systems and Workflow Specialist focused on developing AI-driven systems and design production workflows. She has professional experience as an AI Specialist at Emaar Misr and brings 3 years of hands-on expertise with ComfyUI, as well as local and closed-source generative AI tools. She combines practical industry experience with academic involvement, having previously served as a Teaching Assistant at Ain Shams University. Additionally, she has worked as a Project Design Coordinator at Diaa Consult.

Notes

Format, pace, and studio expectations.

Duration

12 weeks, 1 day per week, 3 hours per session.

Total Time

36 guided hours across a progressive hands-on curriculum.

Course Note

Topics and duration may be modified by the instructor based on participant knowledge and skill level.

Sessions

Twelve sessions from first principles to finished systems.

Sessions / 02

Session01

AI Foundations & ComfyUI setup

AI terminology and useful websites
What is ComfyUI?
Download and install ComfyUI

Session02

Comfy interface & Basic Workflows

Interface overview
Simple text-to-image workflow
Image-to-image workflow

Session03

Prompting & Models

Prompting instructions
Ollama, Florence, Qwen VL, AI Studio
Model types: SDXL, Flux, Qwen Image, Z-Image

Session04

ControlNet & LoRA Workflows

Overview of ControlNet and IP Adapter
Using LoRAs
LoRA training

Task: Train your own LoRA

Session05

Inpainting & Design Iteration

Inpainting workflows
Image editing models
Differences between normal and edit-specific inpainting

Task: Update materials, context, people, objects, and atmosphere

Session06

Enhancement Pipeline

Segmentation and autodetection
Enhancing full images and selected parts
Upscaling

Task: Execute a full image enhancement pipeline

Session07

Krita Integration

Photo manipulation in Photoshop
Intro to Krita and AI plugins
Generate, upscale, and organize presets
Insert ComfyUI workflows and custom parameters in Krita

Task: Deploy a custom ComfyUI workflow within Krita

Session08

3D Generation

3D generation in Hunyuan
Trellis with textures
3D models to segments
Advanced 3D workflows

Task: Construct a textured 3D asset from a 2D design

Session09

AI Video Generation

Intro to local AI video generation
Online video pipeline

Task: Generate a professional cinematic animation

Session10

LLMs, MCP & Automation

Anything LLM
MCP
Use ComfyUI as one tool among many
Website and app workflows with AI tools
Final project setup

Task: The final project

Session11

Vibe Coding, Websites & AI Agents

Intro to vibe coding
Skills and AI agents
Building websites and apps with AI coding tools
UI tools: OpenCode, Claude, VS Code, Antigravity, Codex
Project follow up

Task: Develop and launch a custom web application

Session12

Arch Viz & Interior AI

AI in architectural visualization techniques
AI in interior design techniques
Final project follow up

Task: Finalize the project

Showcase

Scroll through the outcomes this course is built to produce.

Showcase / 03

Multi angles

One scene, many camera decisions.

Temporary Unreal-style placeholder: one environment explored through hero, eye-level, overhead, and detail shots so the output reads like a deliberate sequence.

Wide, mid, and close storytelling frames.
Consistent atmosphere across all views.
Ready for final render replacements later.

Camera setSequencePlaceholder

Unreal-style preview

Angle pack for one concept scene.

Replace with storyboard stills, render crops, or viewport captures.

Suggested future content: one project presented from four curated viewpoints.

Mood board

Atmosphere before geometry.

A previsualization board for lighting, palette, texture, and emotional direction before the final scene gets locked.

Reference logic tied to one visual outcome.
Multiple moods from the same brief.
Strong bridge from inspiration to production.

Look devPaletteAtmosphere

Reference board

Palette, materials, tone, and lighting notes.

Ideal for a collage of references and early style frames.

Suggested future content: one mood board directly connected to the final render set.

Space Perception

Make scale and depth feel believable.

Show how camera height, lens feel, layering, and human-scale anchors transform flat outputs into convincing spaces.

Before-and-after perception fixes.
Lens, horizon, and parallax cues.
Useful for architecture and interiors.

DepthLens feelScale

Spatial study

Foreground, midground, and perspective correction.

Swap in final scene comparisons when ready.

Suggested future content: a perception correction sequence from flat image to immersive space.

Add people

Inject life without breaking the scene.

Context figures added with believable scale, gesture, and lighting so the environment feels inhabited and readable.

Occupancy and circulation cues.
Stronger understanding of function and scale.
Designed as communication, not decoration.

Human contextScene lifeScale

Population pass

Empty composition turned into a socially readable environment.

Best future slot for before-and-after comparisons.

Suggested future content: empty render transformed by contextual population.

3d integration

Blend generated imagery with structured 3D workflows.

A 3D base scene pushed into richer visual outcomes while preserving geometry, intent, and spatial control.

Geometry, segmentation, and AI enhancement together.
2D and 3D as one workflow.
Applicable to architecture, products, and environments.

2D + 3DBridgeWorkflow

Integration sample

Simple mesh, stronger atmosphere, retained structure.

Use later for segmented views, texture passes, and finals.

Suggested future content: 3D base scene translated into a polished presentation frame.

Restore details

Recover resolution, edges, and surface fidelity.

Enhancement workflows for facades, materials, products, and close-up details after generation and upscaling.

Selective restoration instead of blunt sharpening.
Recover texture, edges, and legibility.
Quality-control pass before delivery.

EnhancementUpscaleDetail

Detail recovery

Blurred source refined into a presentation-ready image.

Ideal future placement for zoomed comparison crops.

Suggested future content: crops that prove restored material and edge quality.

Refine people

Fix faces, hands, posture, and integration.

Cleanup workflows for AI-generated characters so they hold up at closer viewing and feel naturally embedded in the final scene.

Anatomy and gesture correction.
Realism without losing mood.
Useful for hero stills and portfolio frames.

Human polishCleanupPost pass

Refinement panel

Close-range cleanup for faces, hands, and figure coherence.

Best future fit: side-by-side evidence of refinement quality.

Suggested future content: portrait or scene crops showing refinement stages.

Style transfer

Hold the structure, shift the visual language.

One brief interpreted through multiple aesthetics while the underlying composition stays coherent and recognizable.

Consistent structure across several looks.
Strategic variation, not random filtering.
Ready for later style comparison boards.

Look shiftVariantsReference style

Style matrix

One composition rendered through multiple visual identities.

Use later for controlled mood and style triptychs.

Suggested future content: one structure shown across several style directions.

Animation

Turn still concepts into motion stories.

Camera moves, parallax motion, and AI video sequences that feel cinematic enough for presentations and launch campaigns.

Still-to-motion workflow.
Atmosphere, timing, and camera path as design tools.
Ready for embedded video later.

MotionVideo pipelineCinematic

Animation preview

Poster frame for a generated motion sequence.

Later replace with a real clip or image sequence.

Suggested future content: one hero frame plus a short motion sample.

Training Lora

Train a visual behavior, not just a single image.

Custom LoRA training presented as a reusable system for repeating a style, identity, material logic, or project-specific direction.

Dataset thinking and training intent.
Repeated consistency across outputs.
Ready for charts and training notes later.

Custom modelConsistencyReusable style

LoRA training story

Dataset in, controlled visual signature out.

Use later for training samples and consistency checks.

Suggested future content: training inputs paired with output consistency.

Live sketch

Move from rough lines to developed imagery fast.

Sketch-to-image workflows where hand-drawn intent survives the jump into rendered atmosphere, materials, and composition.

Fast ideation without losing authorship.
AI as amplification of the designer’s linework.
Good fit for live workshop moments.

Sketch to imageFast iterationAuthorship

Live ideation

Concept sketch expanded into a cinematic design frame.

Best future use: rough sketch on one side, refined output on the other.

Suggested future content: a live sketch progression from linework to polished render.

All in one click

Bundle the workflow into a single repeatable action.

A packaged system where setup, prompting, enhancement, and export are orchestrated as one repeatable production flow.

Complex chains simplified for teams and studios.
Operational rather than experimental.
Ready for workflow diagrams or automation screenshots.

AutomationSystemizedProduction ready

Workflow package

One action orchestrating a full visual production pass.

Later replace with a node graph, UI screenshot, or export sheet.

Suggested future content: compressed workflow overview plus resulting outputs.

vibe coding

Translate the visual thinking into a working product.

The final slot extends beyond images into websites, tools, and product concepts built with AI coding flows and strong visual direction.

Connect visual prompts to real implementation.
AI-assisted product making as part of the same stack.
Ideal home for landing pages, tools, and microsites.

Web outputAI codingProduct thinking

Build preview

From concept language to a shipped interface direction.

Later place screenshots, product flows, or web demos built during the course.

Suggested future content: a live product or website case study built from the course workflow.

Multi angles

Mood board

Space Perception

Add people

3d integration

Restore details

Refine people

Style transfer

Animation

Training Lora

Live sketch

All in one click

vibe coding

Enroll

Reserve your seat.

Early Bird

Take your seat before the first 10 spots are gone.

15% Off

First 10 Seats

Join a hands-on cohort built for designers, architects, visualizers, and artists who want guided practice, repeatable systems, and outcomes worth shipping.

Round Date May 10, 2026

Duration 12 weeks

Guided Time 36 guided hours

Course Language Arabic

Practical workflow training Real-world project outcomes Serious cohort only

Seat Reservation

Secure the early cohort rate.

Standard pricing returns after the first 10 seats are filled.

6,799 EGP

7,999 EGP

12 guided sessions across foundations, workflows, prompting, and production.
Hands-on coverage of ComfyUI, ControlNet, LoRA workflows, inpainting, and enhancement pipelines.
Practical learning path focused on real creative and architectural AI outputs.

Early enrollment pricing is reserved for the first ten participants.

Complete a short guided application and we will contact you on email or WhatsApp to confirm the next step.

Request Details

Questions before applying? Message us directly on WhatsApp.