It was August 2021 when I first joined the PhD program at UC. Apart from face masks, all grad students were talking about Python, programming, and machine learning—and I had no idea what they meant.

Coming from traditional geotechnical field work with a heavily experimental research background, I thought civil engineering was purely classical science. I mean, how long have we been using concrete and steel? But I was about to be amazed at how this “classical” field could embrace such “cool” new tools!

One day, I was discussing research with a friend in my group. He told me he was implementing genetic algorithms to develop settlement prediction equations. Wait, genetic what?

Four years later, I’ve published several peer-reviewed articles on machine learning (ML) applications in geotechnical engineering. I’m now a journal reviewer for several publications in this field and chair of the AI paper review session at our profession’s largest event—the Geotechnical Frontiers 2025.

So what happened between then and now?

Simply put, I discovered the opportunities that ML brings to geotechnical challenges. Coming from a fieldwork background, I understood the challenges geotechnical engineers face in practice.

With these new “cool” machine learning tools, I saw unlimited possibilities to offer solutions—from data exploration to predictive modeling, we can make the data speak to offer insights for better geotechnical practice!

But wait, what is Machine Learning?

Simply, it’s a function that takes data as input and provides an output. In between, it learns the best patterns that lead to this outcome.

For example, with historical earthquake data showing which areas experienced liquefaction, we can use factors like earthquake magnitude, effective stress, and water table depth to predict liquefaction risk in new areas. Whether it’s simple linear regression or advanced neural networks, all ML algorithms essentially do this pattern recognition.

Why Should Geotechs Care?

Why not? I see machine learning as an amazing tool with unlimited possibilities to empower geotechnical engineers.

It helps us gain better insights from existing data, combine it with geotechnical expertise, and optimize our solutions. From site characterization to soil behavior prediction and slope stability analysis—researchers have been developing these applications for decades. Now it’s time for the industry to adopt them.

Will AI Replace Geotechnical Engineers?

At GeoCongress 2023, a speaker said, “AI won’t replace you, but an engineer who knows AI will.” I respectfully disagree! Geotechnical engineering requires specialized skills that will always be valuable, regardless of AI knowledge.

However, AI is pushing our profession’s boundaries, and it’s developing fast. We need to adapt our methods for more robust geotechnical practices.

We can either resist change or adapt and thrive. I choose the latter.

This newsletter is my way of sharing what I’ve learned and helping you do the same. Each week, I’ll break down practical ways ML and AI can make a real difference in geotechnical engineering.

I’d love for you to join us, let’s learn and grow together.