A worldwide postgraduate program at the intersection of AI world models and architectural practice. 18 weeks, 9 modules (F0–F8), real projects from week 1. First cohort Q4 2026. For architects, engineers, designers, and AI/software developers who want to master spatial AI.
MIAW is for architects, software developers, solution architects, designers, engineers, platform architects, technical architects, interior designers, and landscape designers who want to integrate AI into every phase of their design process. It is not a course about AI tools — it is a program about how architectural practice changes when AI is integrated in every phase, from site analysis to documentation, from concept to fabrication.
Duration: 18 weeks at 8–10 hours per week. Compatible with active professional practice. Language: English. Format: Online-first, project-based, real deliverables from week 1. Cohort size: maximum 30 participants. First cohort targeting Q4 2026.
The module starts with a guiding question: what can AI do that is useful for your specific practice, and what can't it?
Deliverable: Configured and documented environment (Claude Code + basic MCP + Git repository), personal knowledge base with at least 10 own documents, and a first functional agent applied to a real task.
The architect's work starts before design: understanding the place, urban context, regulatory constraints, and site opportunities. This module is about how AI transforms that reading and analysis phase.
Deliverable: Automated territorial analysis pipeline applied to a real site of the student, with AI-generated interpretive report and 3D context model integrated in the design environment.
The module works the step from the physical world or a reference to useful 3D geometry, and how to generate or modify geometry from natural language.
Deliverable: 3D capture model of a real space integrated in the design environment, and a natural language-assisted modeling exercise applied to an element of the student's project.
This module works the core of the design process: exploring alternatives, generating variants, and evaluating options against multiple criteria.
Deliverable: Parametric system co-developed with AI for a real element of the student's project, with generative exploration of alternatives evaluated against own criteria and visual synthesis of results.
Many high-impact design decisions are made without sufficient data on building behavior. AI enables that data to arrive earlier, with less technical friction and in an interpretable format.
Deliverable: Simulation pipeline applied to at least two relevant domains for the student's project, with AI-generated results report and design recommendations.
The module works the production of project images, video, and real-time immersive environments with AI integrated in the visualization pipeline. Three levels are worked:
Deliverable: Immersive environments, videos, and images of a real project at all three levels, with visual coherence between them.
An architect produces documentation in all project phases: descriptive reports, technical specifications, licensing files, meeting minutes, construction reports, delivery documentation. Most is written from scratch each time, from data that already exists in the model, notes, or client conversations. This module works how AI automates that production.
Deliverable: Descriptive report, automated specifications of at least two building systems, and regulatory verification report generated with AI from project data, plus demonstration of conversational query to the BIM model.
This module works the transition from digital model to physical component when AI is an active part of the fabrication process. Not learning robotics as an isolated discipline, but understanding how the constructive detail changes when geometry, tool, movement, and assembly are integrated from the start in an intelligent parametric system.
Deliverable: Development of a parametric detail or system oriented to advanced fabrication. Machining and assembly simulation with AI integration. Documentation of the design-fabrication process and decision analysis.
The final module introduces no new tools. It works how to connect everything built in previous modules into a coherent working ecosystem, adapted to the student's specific practice.
Deliverable: Project developed with AI integrated in process stages. Documentation of the working system: agents, knowledge base, flows and adaptive prompts. Presentation of the project and its process.
Gaussian Splatting for 3D scene reconstruction. Large Language Models for design reasoning and documentation. MCP (Model Context Protocol) for connecting AI to AEC tools, files, and databases. Parametric design with AI assistance using Rhino and Grasshopper. Building simulation pipelines with AI-generated interpretation. BIM documentation automation using LLMs. PyTorch and HuggingFace for custom model fine-tuning. Robotic fabrication with AI integration. Computer vision for site monitoring. Graph neural networks for structural analysis. Diffusion models for generative design.
You need either: professional experience in AEC (architecture, engineering, construction) with curiosity about AI, or a software and AI background with interest in physical world applications. You do not need both — MIAW bridges the two domains.
Basic Python is strongly recommended. The program teaches what is needed, but arriving with Python fundamentals allows focus on the domain rather than syntax.
AEC professionals: architects, structural engineers, interior designers, landscape architects, urban planners, BIM managers, construction managers, MEP engineers. AI/software professionals: machine learning engineers, software architects, data scientists, computer vision engineers, robotics engineers with interest in spatial systems.
The Architecture, Engineering, and Construction sector is undergoing its largest transformation since the introduction of CAD. AI world models — systems that build internal representations of physical environments — enable automated code compliance checking, generative structural optimization, construction site monitoring, predictive building maintenance, and autonomous design iteration at scale.
The gap: thousands of architects want to build AI tools. Thousands of AI engineers want to work on physical systems. Almost no one has both. MIAW closes that gap.
Basic Python is strongly recommended. We teach what you need within the program, but arriving with Python fundamentals (functions, loops, basic libraries) will let you focus on the domain rather than syntax.
Yes. Applications involve a short portfolio or work sample, a written statement about why you're applying, and a brief interview. We are keeping cohorts small to maintain quality.
Professional experience in AEC with curiosity about AI, OR software/AI background with interest in physical world applications. You do not need both.
Plan for 8-10 hours per week over 18 weeks. Structured for working professionals — compatible with active professional practice.
Online-first. There will be 1-2 optional in-person intensive sessions per cohort. Optional but strongly recommended.
English. All materials, sessions, and submissions are in English.
Nine progressive modules (F0–F8). F0 builds the foundation: mindset, technical environment, and AI reasoning capacity. F1–F7 work each phase of the design process with AI integrated: reading territory, capturing, modeling, designing, simulating, communicating, documenting, and fabricating. F8 synthesizes everything into your own adaptive design system.
It is structured by phases of the design process, not by tools. Your real project is the working material, not fictional case studies. Each module produces a deliverable directly useful for your professional practice. And it's a gateway to MIAW Advanced for profiles who want to continue in applied research.
Pricing is being finalized. Early applicants will have access to founder pricing. Range being considered: €4,000-8,000 total, with payment plans available.
Targeting Q4 2026. Exact date depends on cohort composition and faculty availability.
Yes. A certificate of completion from MIAW, with project portfolio as primary evidence of learning.
Python, Jupyter, Claude Code, MCP (Model Context Protocol), Rhino/Grasshopper, Revit API (basics), standard ML libraries (PyTorch, HuggingFace), Anthropic/OpenAI APIs, GitHub. Tools for 3D capture (Gaussian Splatting), simulation, and fabrication depending on the student's project. All tools either free or provided as part of the program.
A modern laptop (Mac or Windows) and a stable internet connection are sufficient for most modules. GPU access for some simulation and rendering tasks can be done via cloud.
Project-based from week 1. Your own real project is the working material throughout the program. Each module produces a concrete deliverable directly applicable to professional practice. Learning happens through doing, not through watching demos.
Ethics is not a standalone module — it is integrated throughout the program. We work with the real limits of AI tools: hallucinations, dataset biases, energy costs, and the responsibility that remains with the architect regardless of how much AI was involved in the process.
Applications for the first cohort open ahead of the Q4 2026 launch. Early applicants receive founder pricing. Range being considered: €4,000–€8,000 total with payment plans available. A certificate of completion is awarded, with project portfolio as primary evidence of learning.
The application process is a conversation, not a form. Navigate to miaw.ai and talk with the AI. It will ask about your practice, your goals, and your AI experience — and give you a genuine sense of fit.
Related pages: About MIAW, Full Curriculum (9 Modules), How to Apply.