Ten progressive modules. Each one advances a phase of the design process with AI integrated — from onboarding through territory, modeling, simulation, communication, documentation, fabrication, and full integration.
10
Modules
18
Weeks
300h
Total load
F0–F9
Progression
Program Structure
The curriculum is structured by phases of the design process — not by tools. F0 maps the full
program and frames the course project. F1 builds the technical foundation: programming,
data structures, and AI environment setup. F2 through F8 each advance one phase of the
architectural workflow with AI integrated end-to-end: reading territory, modeling, prototyping, simulating,
communicating, documenting, and fabricating. F9 synthesises everything into the student's
own adaptive AI ecosystem, culminating in a defended final project.
The student's own real project is the working material throughout the program — not fictional case studies.
Each module produces a concrete deliverable directly applicable to professional practice.
Kickoff fortnight. Walk through the program structure, locate the three workshops, and frame the course project so every student arrives at the first module with their question already formulated. This is not a gentle introduction — it's where you decide what you'll be defending in July.
Topics covered: Program structure, cross-cutting workflows, Final Course Project framing, mentor assignment, statement of interests.
Deliverable F0: Map the full program, set up your environment, and define the course project question you will defend in July.
Basic programming, data structures, and AI environment setup. No flashy demos here — this is where you build the foundation you'll use for five months. Skip this module and you'll pay compound interest for the rest of the course.
Topics covered: Programming fundamentals, data structures, databases, data scraping, AI concepts, AI environment setup, MCPs.
Deliverable F1: AI ENVIRONMENT SETUP — Complete configuration of AI tooling, data structures, and file formats for architectural practice, establishing the programming foundation and workflow connections that every subsequent module builds on.
Territory understood as data: topography, landscape, city. The module teaches you to read and produce geographic information with notebooks and models — not to buy layers. The deliverable: build your own territorial layer without depending on a third-party viewer.
Topics covered: GIS, geolocated data display, language model notebooks, programming 2, topography scraping, landscape modeling, smart city modelling, MCPs 2.
Deliverable F2: TERRAIN GENERATION TOOLKIT — Procedural and bespoke system for generating and modifying digital terrain using computational rules, geospatial data and simulation feedback loops, enabling controllable landscape formation and iterative spatial design.
Create, modify, and control geometry with AI. From meshes, NURBS, and SubDs to fast-modelling workflows and topology aids — the module builds the toolkit for generating and adapting complex forms at design speed.
Topics covered: Meshes, NURBS, SubDs, fast modelling techniques, topology and modelling aids.
Deliverable F3: AI ASSISTED MODELING AGENT AND PIPELINE — End-to-end AI-assisted architectural modeling workflow that transforms multimodal conceptual inputs (text, sketches, references or spatial constraints) into usable 3D models through generative AI.
This is where you decide how to prototype for the rest of the program. Visual programming, assisted scripting, parametric vs. code. The module forces you to choose tool and method for each situation, rather than defaulting to what you already know.
Topics covered: Visual programming, advanced prototypes, AI-assisted scripting, parametric vs code, topologies vs typologies, procedural modelling, deterministic vs stochastic, rules as tools.
Deliverable F4: ADAPTIVE RULE-BASED TYPOLOGY GENERATOR — Computational design system for generating adaptive architectural typologies based on rule-based logic, where spatial configurations emerge from encoded constraints, behavioural rules and environmental parameters, enabling continuous generation, variation and optimisation of building typologies in response to changing contextual and performance conditions.
Multidisciplinary simulation with AI: structure, environment, flows, optimization. The emphasis is on limits — what a network simulates well, what a classical FEM does better, when a multi-agent system is the right call. Learning to validate results is half the module.
Topics covered: Simulation types and cost, smart workflows, optimization algorithms, AI result validation, alternative simulations, finite element models, ML / EVO / EMO / AI, multiagent systems.
Deliverable F5: PREDICTIVE PERFORMANCE BUILDING MODEL — Integrated computational building system that combines simulation, AI-based prediction and real-time performance analysis to anticipate structural, environmental and operational behaviour, enabling the design of buildings whose geometry and systems are continuously evaluated.
Image and video production with generative models, return from audiovisual to 3D model, immersive environments. The module's question is not how to generate a pretty render, but how to build a coherent visual language for an AI project applied to the built environment.
Topics covered: AI image tooling, AI image models, model to image, image to video, video and image to model, image to world.
Deliverable F6: AI IMMERSIVE APPLICATION — AI-assisted multimedia production and digital immersion framework that transforms architectural models and simulation outputs into coherent visual narratives, combining image, video and real-time 3D environments to communicate design, performance data and spatial experience through immersive and interactive interfaces.
What is not documented does not exist on site. The module works through advanced BIM, entities, exchange formats, and AI-assisted drafting. It also covers document management and the boundary with robotics and construction.
Topics covered: Advanced BIM entities, data management and formats, details, materials, articulations, construction constraints and robotics 101, BIM super-integrations, AI-driven drafting.
Deliverable F7: CLOUD AI-BIM DOCUMENTATION PLATFORM — Cloud-based openBIM system for automated architectural documentation and project information management, enabling real-time generation, coordination and versioning of documentation directly from BIM models, integrating interoperable openBIM standards, AI-assisted content generation and collaborative data workflows across design, engineering and construction environments.
Advanced digital fabrication, robotics, and complex assemblies. The module closes the digital–physical–digital loop: the model is built, measured, and returned to the system corrected. Optimization workflows, gcode, and smart fabrication strategies designed for real-world production.
Topics covered: From digital to physical and back, smart workflows, optimization for digital fabrication, robotics 101, 3D printing, gcodes and paths.
Deliverable F8: AI DIGITAL FABRICATION ALGORITHM SCRIPTING FOR COMPLEX GEOMETRY — AI-driven computational fabrication framework for the design and production of singular architectural components, where algorithmic geometry generation, material constraints and robotic manufacturing processes converge to produce unique, non-repetitive building elements optimized through simulation, fabrication feedback and construction-aware design logic.
The closing module adds no new topics — it connects them. Each student integrates tools, data, models, and workflows from the previous nine modules into a personal ecosystem they'll defend before the final jury.
Topics covered: Multimodal integration, bespoke AI workflows, Final Course Project, final jury.
Deliverable F9: GLOBAL AI INTERACTIVE INTERFACE — A bespoke, adaptive multimodal AI ecosystem integrating all tools, data, and workflows from every previous module, ready to defend.
Prerequisites
Technical
Basic Python is strongly recommended. The program teaches what you need within it, but arriving with Python fundamentals (functions, loops, basic libraries) lets you focus on the domain, not syntax.
Professional
Professional experience in AEC or software/AI development background. You do not need both — MIAWS bridges them.
Hardware
A modern laptop (Mac or Windows) and stable internet connection. GPU access for simulation tasks is available via cloud — not required locally.
Questions about the curriculum?
Ask the AI — it can explain any module, discuss prerequisites, or help you understand how the program fits your background.
