01
Observe
Experimental evidence
I combine mechanical, durability, thermal, mineralogical, electrical, and microscopic measurements to determine how alternative constituents reshape cementitious systems.
PhD Student and Graduate Research Assistant · SMARIN Lab
Mayaz Uddin Gazi
Sustainable cementitious materials, Advanced Concrete, and ML/Computer Vision
I am a PhD candidate in Civil and Architectural Engineering at the University of Miami, working in the SMARIN Lab under Dr. Xianming Shi. My work sits Between the materials laboratory and the computer, I study how cementitious systems behave when conventional constituents are replaced by supplementary and alternative materials, and I build computational models that make that behavior predictable. Experimentally, I work across fresh properties, mechanical performance, durability, and microstructure, using TGA, XRD, XRF, SEM-EDS, isothermal calorimetry, electrical resistivity and impedance spectroscopy, shrinkage, and leaching methods under ASTM, AASHTO, and ACI protocols. Computationally, I work in physics-informed and physics-constrained machine learning, Bayesian optimization for mixture design, graph and deep neural networks, and computer vision for microstructural image analysis. I am most interested in problems where neither the chemistry nor the algorithm is Sufficient on its own.
- Institution
- University of Miami
- Department
- Department of Civil and Architectural Engineering
- Advisor
- Prof. Xianming Shi
Research philosophy
From material evidence to responsible prediction.
My work follows a simple research logic: observe how a material behaves, explain that behavior with engineering science, and translate the evidence into models that remain useful beyond a single dataset.
02
Understand
Physics-aware intelligence
I connect hydration, pore structure, age, and performance through constrained and uncertainty-aware learning, with particular attention to behavior outside the training domain.
03
Translate
Engineering decisions
I use prediction and optimization to support realistic decisions about mixture design, durability, embodied carbon, and cost while respecting material and specification limits.
Research agenda
Current research themes
Questions that connect sustainable construction materials with physically credible computation.
Biochar-modified cementitious materials
Internal curing, pore refinement, carbon storage, mechanical performance, and dimensional stability in biochar-blended mortar.
Seawater and sea-sand binder systems
Hydration, phase assemblage, resistivity, transport, and durability of fly-ash and biochar mortars produced with alternative water and aggregate sources.
Physics-informed machine learning
Hybrid learning frameworks that embed maturity, hydration, and engineering constraints to improve physical consistency and extrapolation.
Digital twins and mix optimization
Data-assisted concrete quality control and multi-objective mix design balancing strength, cost, and embodied carbon.
Microstructure image analysis
Reproducible analysis of cementitious microstructures using microscopy, phase mapping, and image segmentation.
In progress
Selected projects
Active experimental and computational work. Status labels distinguish projects from peer-reviewed publications.
Seawater–fly ash–biochar mortar
Experimental study of strength, electrical response, shrinkage, hydration products, and contaminant immobilization across curing age.
Physics integration in concrete strength learning
Comparison of conventional regression, neural networks, and hard-constrained physics-informed models under interpolation and material-domain shift.
Sustainable mortar digital twin
A multi-objective decision framework connecting early-age measurements with strength, cost, and embodied-carbon optimization.
Machine learning for alkali-activated concrete
Literature-derived modeling with source-aware validation to test whether strength models generalize to unseen studies.
Research outputs
Journal publications
-
01Published2026Journal article
HEStacked-FSML: Data-Efficient Prediction of Nano-Modified Concrete Compressive Strength
Journal of Computing in Civil Engineering, 40(6), 04026084 · ASCE
doi:10.1061/JCCEE5.CPENG-7508 -
02Published2025Journal article
Few-Shot Meta-Learning for Concrete Strength Prediction: A Model-Agnostic Approach with SHAP Analysis
AI in Civil Engineering, 4(1), 20 · Springer Nature
doi:10.1007/s43503-025-00064-8 -
03Under reviewYear pendingJournal manuscript
Predictive Modeling of Water Pollution in the Surma River Using the Water Pollution Index Framework
Scholarly communication
Conference papers
Peer-reviewed proceedings and accepted conference contributions.
2025
Published
Enhancing Concrete Performance: Machine Learning Models for Predicting and Classifying Tensile Strength
IEEE International Conference on Electrical, Computer and Communication Engineering (ECCE 2025)
doi:10.1109/ECCE64574.2025.110139912025
Published
Data-Driven and Physics-Informed Modeling of Rice Husk Ash Concrete for Strength and Sustainability Optimization
28th International Conference on Computer and Information Technology (ICCIT 2025), IEEE
2025
Published
Deep Learning Framework for Accurate and Scalable Landslide Hazard Assessment
28th International Conference on Computer and Information Technology (ICCIT 2025), IEEE
2025
Published
AI-Based Urban Risk Modeling: Leveraging Physics-Informed Neural Networks for Real-Time Structural Stress Simulation of a Structure
28th International Conference on Computer and Information Technology (ICCIT 2025), IEEE
2024
Published
CrackClassification4: Classifying Four Types of Cracks in Structural Surfaces Using CNNs
27th International Conference on Computer and Information Technology (ICCIT 2024), IEEE
doi:10.1109/ICCIT64611.2024.110225742025
Accepted
Meta-Learning Model for Low-Data River Water Quality Analysis in Northeastern Bangladesh
16th IEEE International Conference on Computing, Communication and Networking Technologies (ICCCNT)
2024
Accepted
Flood Prediction and Forecasting Using ANNs and a Fuzzy Logic Model in Sylhet, Bangladesh
7th International Conference on Advances in Civil Engineering (ICACE), CUET - [Springer]
Background
Experience and education
2026–present
Graduate Research Assistant
SMARIN Lab, University of Miami
Experimental characterization of sustainable cementitious materials and development of physics-informed learning frameworks under Prof. Xianming Shi.
2026–present
Graduate Teaching Assistant
University of Miami
Support for undergraduate civil engineering instruction, laboratory activities, assessment, and student learning.
2025
AI Engineer
DebuggerLab
Developed, evaluated, and communicated machine-learning workflows for engineering applications.
January–June 2025
Research Assistant
Sylhet Engineering College
Worked on experimental design, data preprocessing, and predictive modeling for concrete research.
2026–present
PhD in Civil and Architectural Engineering
University of Miami
Coral Gables, Florida · Advisor: Prof. Xianming Shi
2020–2025
BSc in Civil Engineering
Sylhet Engineering College, Shahjalal University of Science and Technology
Sylhet, Bangladesh · CGPA: 3.40/4.00
Capabilities
Methods and tools
Experimental
- Compressive and split-tensile strength
- Shrinkage and dimensional stability
- Bulk resistivity and EIS
- Sorptivity and chloride transport
- Leaching and contaminant immobilization
Materials characterization
- XRD phase analysis
- TGA and DTG
- SEM, BSE, and EDS mapping
- Isothermal calorimetry
- Pore-structure interpretation
Data and modeling
- Python, MATLAB, and Jupyter
- PyTorch and scikit-learn
- Physics-informed neural networks
- Source-aware and out-of-domain validation
- SHAP and uncertainty analysis
Engineering workflows
- COMSOL Multiphysics and finite elements
- Life-cycle assessment with openLCA
- Bayesian optimization
- Git and reproducible research
- Technical and academic communication
Leadership
Service and engagement
2024–2025
President
CE Center of Research
Led student research initiatives, seminars, workshops, and peer learning activities.
2024–2025
Vice President
SEC Civil Society
Coordinated academic and professional-development activities for civil engineering students.
Research notes
Blog
Short notes on methods, evidence, reproducibility, and lessons from work in progress.
Building a Research Portfolio Around Evidence
Why I am organizing my experimental, computational, and publication work in one transparent research record.
Read note →Contact
Let’s discuss sustainable materials and engineering AI.
Open to research collaboration, scholarly exchange, and work connecting cementitious-material experiments with reliable data-driven methods.
Email Mayaz