Training for architects, urban planners, engineers and project owners

The Architecture & Intelligences Research Chair (EVCAU Laboratory / ENSA Paris-Val de Seine) offers you the opportunity to master the creative development and production methods and tools that are transforming architecture, urban planning and design. Generative AI, digital twins, data science, computational design, GIS, AI agents, the AI Act, reuse and predictive design: in 1 to 5 days, move from simple curiosity to hands-on practice with concrete, distinctive case studies drawn from the trainers' and architects' own work:  exploring design variants, diagnosing a territory, automating practice tasks, or designing with reused materials.

Research-based training focused on design projects, for more creative, productive and intelligent practices.

Architecture & Intelligences — 8 short courses to design differently

The Architecture & Intelligences Research Chair (ENSA Paris-Val de Seine, EVCAU Laboratory) offers 8 short courses lasting 1 to 5 days, 22 days in total. They are designed for professionals in architecture, urban planning, design and project ownership who wish to develop their skills or explore a new field of practice before undertaking longer training (specialised master's, PhD, etc.).

Each course is linked to one of the Chair's four research areas and is taught by university lecturers and researchers (associate professors or professors) or professional trainers. Several professional courses with a strong technical component (F4, F6, F7) can be delivered by a Qualiopi-certified training organisation, making them eligible for OPCO and France Travail funding.

Recommended pathways:

• Practices getting started with AI: F1, then F7, then F6.
• Territory and data: F4, then F2, then F3.
• Low-carbon design: F5, then F8.

F1 — Generative AI for architectural design

Duration: 2 days (14 h) · Format: in person, 12 participants maximum · Research area: Augmented design

Audience: architects, landscape architects, designers, project managers in architectural practices.

Prerequisites: no coding skills required. Each participant brings a laptop and, if possible, an ongoing project.

Objectives. By the end of the course, participants will be able to:

1. Choose the right generative AI tool for each project phase (text, image, 3D, video).
2. Write structured prompts for the brief, the sketch design, the atmosphere and the design statement.
3. Critically assess AI outputs: biases, formal clichés, constructive consistency.
4. Identify legal risks: copyright, confidentiality of client data, mandatory disclosures.

Programme.

Methods: 30% theory, 70% hands-on computer practice. Examples are drawn from real architectural projects from the work of the trainers or of the participants.

Deliverables: a reusable practice prompt book and a board of design variants for the participant's project.

Assessment: graded practical exercise at the end of Day 2 and a 15-question multiple-choice quiz.

F2 — AI- and GIS-augmented territorial diagnosis

Duration: 3 days (21 h) · Format: in person or hybrid, 15 participants maximum · Research area: Resources & ecosystems

Audience: urban planners, architects, local authority staff, environmental consultancies, developers.

Prerequisites: basic spreadsheet skills. No GIS knowledge required.

Objectives. By the end of the course, participants will be able to:

1. Structure a territorial diagnosis using a systemic approach (ecological, social, economic and spatial components).
2. Find and cross-reference relevant open data in QGIS.
3. Build vulnerability indicators and map them.
4. Use AI-assisted classification to read land use and urban dynamics.
5. Present a clear diagnosis, with scenarios and recommendations.

Programme.

Methods: supervised hands-on work with prepared datasets. The running case study is a municipality in the Île-de-France region.

Deliverables: an atlas of 5 to 8 maps and a systemic diagram of the territory's strengths and weaknesses.

Assessment: 50% for practical work, 50% for the mini-diagnosis presentation.

F3 — Digital twins and BIM for the sustainable city

Duration: 3 days (21 h) · Format: in person, 12 participants maximum · Research area: Augmented design

Audience: BIM managers, architects, public and private project owners, property and asset managers.

Prerequisites: basic knowledge of BIM (digital model, IFC). Experience with BIM software (Revit, Archicad) is desirable.

Objectives. By the end of the course, participants will be able to:

1. Distinguish between BIM models, CIM and digital twins, and understand what decisions each one supports.
2. Apply the principles of ISO 19650 and open exchange formats (IFC, CityGML, CityJSON).
3. Assemble a simple block-scale digital twin connected to dynamic data (energy, sensors).
4. Use simulations (energy, comfort, use) to inform collective decision-making.

Programme.

Methods: demonstrations, followed by guided practice on a supplied model. Participants may also work on their own organisation's model.

Deliverables: a block-scale digital twin and a 2-page scenario comparison note.

Assessment: continuous assessment of exercises and a final presentation.

F4 — Python and data science for architects and urban planners

Duration: 5 days (35 h) · Format: in person or live online, 12 participants maximum · Research area: AI for architecture

Audience: architects, urban planners, landscape architects and project officers who want to process their own data.

Prerequisites: no programming experience required. Spreadsheet logic is enough.

Objectives. By the end of the course, participants will be able to:

1. Write simple Python scripts in Jupyter to automate repetitive tasks.
2. Clean, combine and analyse tabular and geographic data (Pandas, GeoPandas).
3. Produce interactive maps and charts for study reports.
4. Train and interpret a first machine learning model applied to the city.
5. Use an AI assistant to write and debug code, while knowing how to check the result.

Programme.

Methods: "one concept, one exercise" teaching approach. All notebooks are provided and remain accessible after the course.

Deliverables: a documented project notebook and a library of reusable functions.

Assessment: daily self-correcting exercises and presentation of the Day 5 project.

Note: this is the course best suited to joint delivery with 2IIA. It is eligible for OPCO and France Travail funding through 2IIA's Qualiopi certification.

F5 — Computational design and generative algorithms

Duration: 3 days (21 h) · Format: in person, 12 participants maximum · Research area: Augmented design

Audience: architects, designers, structural and façade engineers.

Prerequisites: experience with a 3D modelling tool (preferably Rhino). No programming experience required.

Objectives. By the end of the course, participants will be able to:

1. Translate an architectural intention into rules and parameters.
2. Build a parametric model in Grasshopper, from geometry to contextual data.
3. Run a multi-objective optimisation (solar, views, materials, floor areas) and read a Pareto front.
4. Couple the generative model with AI to quickly sort and evaluate thousands of variants.

Programme.

Methods: short theory sessions, followed by intensive practice. All scripts are provided and annotated.

Deliverables: a documented parametric script and a board of optimised variants.

Assessment: quality and readability of the script, relevance of the optimisation criteria.

F6 — Data, ethics and the AI Act for the city

Duration: 1 day (7 h) · Format: in person or online, 20 participants maximum · Research area: cross-cutting

Audience: elected officials and local authority directors, heads of architectural practices, developers, project owners, IT directors.

Prerequisites: none.

Objectives. By the end of the course, participants will be able to:

1. Determine the risk level of an urban AI system under the AI Act.
2. Know the obligations that already apply and the timetable for those to come.
3. Combine the AI Act and the GDPR for projects based on sensors, cameras or personal data.
4. Lay the foundations of data governance: openness, sovereignty, digital sobriety.

Regulatory update included in the programme. Regulation (EU) 2026/1744, known as the "AI digital omnibus", entered into force on 27 July 2026. It postpones the obligations for high-risk systems under Annex III to 2 December 2027, and those under Annex I to 2 August 2028. Obligations already in force remain applicable, in particular prohibited practices and AI literacy training for staff. 

Programme.

Methods: short presentations, real case studies and an assessment grid given to each participant. The regulatory content is updated before each session, as the framework is evolving rapidly.

Deliverables: an AI Act / GDPR self-assessment grid for an urban project and a model AI usage charter.

Assessment: 20-question multiple-choice quiz and workshop presentation.

F7 — AI agents and automation in architectural practice

Duration: 2 days (14 h) · Format: in person, 12 participants maximum · Research area: AI for architecture

Audience: partners and heads of architectural practices, project managers, administrative managers, quantity surveyors.

Prerequisites: regular use of a conversational assistant (ChatGPT, Claude, Mistral). F1 is recommended but not mandatory.

Objectives. By the end of the course, participants will be able to:

1. Identify the practice's tasks that can be automated and estimate the time saved.
2. Build a document assistant (RAG) on the practice's documents: DTU technical standards, local urban plans (PLU), standard specifications (CCTP), project archives.
3. Configure a simple agent that chains several steps, with human validation.
4. Put in place confidentiality and security rules suited to client data.
5. Write an adoption roadmap for the practice.

Programme.

Methods: each participant works on a real case from their own practice, with anonymised documents if needed.

Deliverables: a working prototype of an internal assistant and a 6-month adoption roadmap.

Assessment: prototype demonstration and roadmap review.

F8 — AI for reuse and the circular economy in construction

Subtitle: searching for, characterising and designing with disparate and limited material stocks.

Duration: 3 days (21 h) · Format: in person, 12 participants maximum, with a visit to a site or a reuse platform · Research area: Resources & ecosystems

Audience: architects, reuse consultants, PEMD auditors (pre-demolition audits), structural engineering firms, project owners, social housing landlords.

Prerequisites: basic knowledge of design projects or construction sites. F5 is useful for Day 3 but not mandatory.

Challenge. Reuse reverses the usual logic of design. Instead of choosing a material from an unlimited catalogue, we work with an existing stock that is heterogeneous, scattered and available in small quantities. Since 1 July 2023, the PEMD audit (products, equipment, materials and waste) has been mandatory before any demolition or major renovation. This audit, filed on the CSTB platform, makes reusable material stocks visible. The extended producer responsibility (EPR) scheme for building products and materials has been operating since 2023. 

AI comes into play at three stages: finding material stocks, defining the resource (characterising it and making it comparable), then designing with it predictively.

Objectives. By the end of the course, participants will be able to:

1. Search for and aggregate material stocks from heterogeneous sources: PEMD audits, reuse platforms, project owners' inventories, site surveys.
2. Structure a consistent resource database from disparate data (photos, PDFs, spreadsheets, 3D surveys).
3. Use AI (computer vision, text extraction, estimation) to characterise elements: dimensions, condition, likely performance, carbon avoided.
4. Design under stock constraints: assign available elements to a project, optimise, and generate variants compatible with the material stock.
5. Anticipate uncertainty: limited quantities, availability over time, need for technical and insurance qualification.

Programme.

Methods: a mix of framing sessions, demonstrations and workshops on a real dataset. This dataset combines a PEMD report, a photographic inventory and a 3D survey. A visit to a reuse platform or a strip-out site is planned, depending on available partnerships.

Deliverables:

• a structured, reusable resource database;
• a design script working under stock constraints;
• a project sheet showing the reuse rate, the carbon avoided and key technical and insurance issues.

Assessment: 40% for the quality of the resource database, 60% for the workshop project and its presentation.

Potential partners: Saint-Gobain and Point P, partners of the Chair, for questions of material take-back and supply chains; AQC, also a partner, for the technical and insurance qualification of reuse; an EPR eco-organisation for building products and materials for material-stock data.

General information

Location: ENSA Paris-Val de Seine, 3-5 quai Panhard et Levassor, Paris 13th arrondissement. Courses F4, F6 and F7 can also be delivered remotely.

Formats: open multi-organisation sessions on the Chair's calendar, or in-house sessions for a practice, a local authority or an industry partner, with a tailored case study.

Assessment and certificate: pre-course positioning, assessment of learning at the end of the session, satisfaction surveys immediately after and 3 months later. Each participant receives a certificate from the Chair.

Trainers: Lecturers, researchers and trainers from the Architecture & Intelligences Research Chair and its partners, such as Pierre Vincent, Roland Dehghan Kamaragi and David Serero.

Accessibility: a disability officer reviews every adaptation request before registration.

Funding: Several courses (e.g. F4, F6, F7) can be delivered jointly with a Qualiopi-certified training organisation, enabling funding through OPCOs, France Travail (AIF) or the employer's skills development plan. 

Indicative fees: a daily rate per participant is offered for open multi-organisation sessions, depending on the technical level. For in-house sessions, a daily group rate is set on quotation.

Ready-to-deliver courses 
We have also created a MOOC (Massive Open Online Course) on artificial intelligence in architecture, available through our partner Aivancity.

For more information: