Zigurat Global Institute of Technology — Group 5 (2025–2026)
Welcome to the official GitHub organization for Group 5. This space hosts our collective research and development projects completed during the MAICEN-1125 Master's program. Our work focuses on the intersection of Artificial Intelligence, Parametric Design, and BIM to solve complex challenges in the AECO industry.
We are a multidisciplinary group of professionals dedicated to integrating AI into the construction lifecycle.
| Name | Primary Specialization | Key Contributions |
|---|---|---|
| Mark Shane Haines | Project Lead / Integration | Repo Owner, Integration Lead, SARIMA & YOLO Logic |
| Letícia Cristovam Clemente | Computational Design / Data | Floor & Level Logic, Prophet Forecasting, Dataset Annotation |
| Malak Yaseen | Structural / Statistics | Column Grid Generation, ACF/PACF Analysis, Model Training |
| Marc Azzam | BIM / Visualisation | Façade Design, Seasonality Analysis, Error Evidence |
| Osama Ata | Governance / BIM Strategy | Rhino.Inside Revit Output, Data Preprocessing, Presentation |
Tools: Grasshopper, Rhino.Inside Revit, Python
A parametric multi-storey building generator that transforms plot boundaries into fully-formed BIM elements (Levels, Floors, Columns, Façades) in Autodesk Revit.
- Key Feature: Real-time BIM synchronization and footprint validation (<60% coverage).
- Outcome: Automated generation of 5–50 storey buildings with live data summaries.
Tools: YOLOv8, Computer Vision, Pseudo-Labeling
An AI-powered safety system that monitors construction sites for PPE compliance (Helmets, Safety Vests, Goggles) and detects violations (Bare Heads).
- Key Feature: Innovative pseudo-labeling pipeline to expand detection classes without manual annotation.
- Result: 91.9% recall on bare-head detection, prioritizing life-safety.
Tools: Python, Prophet, SARIMA, Statsmodels
End-to-end analysis and forecasting of 16 years of hourly energy consumption data (PJM West) to support energy procurement in the AECO sector.
- Key Feature: Multi-scale seasonality analysis and comparison between Meta's Prophet and SARIMA models.
- Result: High-accuracy forecasting with a 6.5% MAPE on a 52-week holdout.
| Domain | Technologies |
|---|---|
| Computational Design | Rhino 7/8, Grasshopper, Rhino.Inside Revit |
| Artificial Intelligence | PyTorch, Ultralytics (YOLOv8), Prophet, Scikit-learn |
| Data Science | Python (Pandas, Numpy, Seaborn, Statsmodels) |
| BIM | Autodesk Revit 2022+ |
MAICEN-1125 is the Master's program in Artificial Intelligence for Architecture and Construction at the Zigurat Global Institute of Technology. Our group (Group 5) focuses on "AI for Project Optimisation, Innovation, and Ethics."
All repositories within this organization are released under the MIT License unless otherwise specified in the individual repository.
"Bridging the gap between algorithmic design and intelligent construction."
Zigurat | Group 5 | 2025–2026