Amir Moayyedi

As a research assistant at Portland State University, I specialize in Deep Learning, Machine Learning, and Computer Vision with applications in structural analysis and agriculture. Leveraging over a decade of programming experience across MATLAB, OpenSees, Python, and SQL, I develop innovative AI solutions that bridge theoretical concepts with practical applications. My doctoral research combines mathematical expertise with advanced machine learning and neural network techniques, demonstrating my commitment to pushing the boundaries of AI integration across diverse domains.

I am continuously expanding my knowledge in artificial intelligence while maintaining a strong foundation in mathematics and computational methods.

-

AI Engineer, Xogrox AI

2024- present

• Annotated grapevine disease images using Roboflow and developed Python code for vineyard disease detection, utilizing different YOLO models (YOLOv8, v9, GELAN, v10, v11) for real-world applications.
• Hyperparameter Tuning using Optuna (optimization framework) to tune hyperparameter into best YOLO mode.
• Developed the Navigation Stack in ROS, Gazebo, and Rviz to facilitate robotic navigation in vineyards (Ongoing).

-

Data Science researcher in Structural Engineering, Portland State University (PSU)

2022- Present

• A Reinforcement Learning-Driven Framework for Decision Tree Transformation: From Soft decision Trees to Linear Model Tree and Hard Decision Trees for Enhanced Interpretability Using Temperature, Pruning and Regularization and other technics (PyTorch).
• Implemented a new probabilistic risk assessment method for bridge networks using Gauss–Legendre Quadrature for uncertainty quantification (Python).
• Applied dynamic programming and Bellman Optimality Equations for policy iteration and value iteration in grid world problems (Python).
• Leveraged data mining for network-level planning with proficiency in benefit-cost analysis using value iteration (MDP) (Python).
• Created a decision model optimizing bridge maintenance, repair, and reconstruction (MR&R) policies, accounting for earthquake-related uncertainties (Python).
• Optimum life-cycle maintenance strategies for deteriorating highway bridges subject to seismic hazard by a hybrid Markov decision (Python).
• Implemented a five-armed bandit problem with greedy and epsilon-greedy action selection algorithms (Python).

-

Data Mining Engineer, Behsazan

2016 - 2021

• Used Probabilistic and statistical models to estimate the probability of hazard occurrence and its potential cost on offshore structures (Open Sees, MATLAB).
• Used statistical techniques to identify patterns,

As a research assistant at Portland State University, I specialize in Deep Learning, Machine Learning, and Computer Vision with applications in structural analysis and agriculture.