Introduction to Artificial Intelligence for Engineers & STEM

This course is designed specifically for engineers and STEM professionals who want a rigorous, application-driven introduction to machine learning. Rather than presenting AI as an abstract or purely software-oriented discipline, the course frames every concept within engineering workflows, physical modeling, and research practice.

Most machine learning courses emphasize generic datasets and business-oriented use cases. In contrast, this course connects core ML methodology directly to engineering problems.

The course begins with Linear Regression Fundamentals, using the concrete example of predicting elastic stress in a steel specimen to introduce essential terminology and concepts, including loss functions, optimizers, and generalization. This establishes a mathematically grounded understanding of supervised learning before moving to more advanced models.

Students then move beyond black-box modeling through Symbolic Regression and Genetic Programming, learning how to discover interpretable, closed-form mathematical relationships directly from data. For example, the course demonstrates how to recover governing-style equations such as predicting the tip deflection of a cantilever beam. This module emphasizes interpretability, physical insight, and equation discovery.

The course also covers Neural Networks and Deep Learning, including the construction and training of fully connected feedforward neural networks (FNNs) and convolutional neural networks (CNNs). These architectures are applied to realistic engineering tasks, such as image-based crack detection in concrete surfaces, illustrating how deep learning supports inspection and structural health monitoring.

Additionally, the course explores the use of LLMs (large language models), such as ChatGPT, Claude, and Gemini, for writing and debugging code. It also examines responsible ways to use these tools, as well as common pitfalls, including hallucinations and other important limitations. The course further introduces agentic coding with LLMs, showing how it can speed up workflows while emphasizing the risks associated with its use.

Finally, a hands-on capstone project allows participants to translate theory into applied skill. Students may choose to work on tabular prediction problems, symbolic regression–based equation discovery, image-based inspection tasks, or surrogate modeling and optimization problems. The capstone is open-ended, encouraging participants to engage with real engineering-style datasets and decision workflows.

By the end of the course, participants will understand not only how machine learning models work, but how to deploy them responsibly and effectively within engineering contexts. The objective is to provide a structured foundation that enables engineers to integrate AI into research, design, and analysis with confidence.