Doctor of Philosophy - PhD, Civil Engineering; Artificial Intelligence in Construction

Grade: 4.46/4.5 • Dissertation topic: “Predicting damages of reinforced concrete panels under impact loads using machine learning” • Research content: Developing novel techniques to deal with data problems using various machine learning algorithms. • Main achievements: 07 published papers (03 in the top 10% of journals)

Master of Science - MS, Structual Optimization; Computer Programming, Specific Applications

Grade: 4.43/4.5 • Thesis topic: “Multi-material structural topology optimization method with elastic buckling constraints” • Research content: Developing effective methods for structural optimization under buckling problems. • Main achievements: 04 published papers (01 in the top 10% of journals)

Bachelor of Engineering - BE, Civil Engineering

Grade: 8.2/10 • Relevant courses: Project management, Cost analysis, Civil planning, Organization of construction process, Work safety, Consulting and Supervision, Water supply, Electrical system.

Data Scientist

• Be the person in charge of a project that assesses the effectiveness of a learning platform created by a large educational corporation of Japan by extracting valuable insights from given data.
• Collaborate with business team to identify and clarify client\’s business challenges and requirements of the project.
• Construct hypotheses, design experiments, and develop appropriate analysis methods to achieve project’s goals.
• Perform exploratory data analysis and data preprocessing.
• Conduct various statistical analyses, regression, correlation analyses, variable impact analysis using statistical methods and machine learning techniques.
• So far achievements:
o Obtained valuable insights from the analysis and almost completed 2nd phase of the project with positive feedback from the client.

Doctoral Researcher

• Performed data pre-processing for training machine learning models to predict structural damages including cleansing, outlier removing, feature selection, data transformation, categorical feature encoding, etc.
• Analyzed data and extracted insights from given data such as distribution, statistical information, correlation coefficient, statistical significance, etc.
• Implemented data dimensionality reduction using PCA and t-SNE techniques.
• Built many impact damage prediction models using various machine learning algorithms such as Artificial Neural Networks, XGBoost, Decision Tree, Random Forest, KNN, SVM, etc.
• Optimized hyperparameters of machine learning models to improve prediction performance by using various optimization algorithms such as Bayesian optimization, Genetic algorithm, Particle swarm optimization, etc.
• Proposed a novel clustering-based data splitting method for the small-imbalance dataset problem to improve damage classification performance using HDBSCAN clustering and stratified sampling method.
• Proposed a new tabular data augmentation algorithm for an imbalance dataset using a combination of the Wasserstein GAN with Gradient Penalty (WGANGP) and Model Agnostic Meta-learning (MAML) algorithms, which was implemented by Tensorflow/Keras framework.
• Developed a hyperparameter optimization strategy for neural network models to improve prediction accuracy when learning from a small dataset.
• Presented research works at many international conferences about advanced technologies in construction.
• Was a teaching assistant in a machine learning class for undergraduate students.

Structural Designer

• Performed structural design for many different types of buildings such as apartment buildings, hotels, and manufacturing factories.
• Modeled and calculated the resistance capacity of structures using MIDAS software for many concrete, steel, and steel-concrete buildings.
• Performed on-site quality assessments and propose reinforcement plans for many schools and hotel buildings across Korea.

Main achievements:
– Created a mini software written by Python that contributed to significantly minimizing the labor cost when modeling a building’s foundation. In particular, it often takes about 20 minutes ~ 1 hour (depending on the building size) to create the initial elements, column force positions, etc… ready for analyzing a building foundation. My mini-software helps to finish this task within a minute regardless of the building size.
– When the Korean government suddenly requires to investigate and evaluate the earthquake resistance capacity of all school buildings across the country, after a magnitude 5.4 earthquake in 2017, damaging many school buildings. My team was stuck in adapting a new earthquake evaluation method required by the government regulation. I have contributed to doing research to understand the method, and then pushed my team moving forward. It helped my ex-company to have a quick adaption and be confident to gain many other similar projects.

Quantity Surveyor

• Collaborated with other people to measure and cross-check the quantity of
various building/structure types such as concrete, steel, formwork, etc.

Main achievement:
– Created a tool to accelerate the measurement cross-checking process which was implemented manually in the excel program. The application helped the company to save much working time, which allowed the company to get many projects at the same time.

Assistant Construction Manager

• Supervised the construction process at the site.
• Assisted the Construction manager to make progress report.