Every day, vast amounts of water are analyzed for pollutants, nutrients, and other chemicals using advanced laboratory techniques. These methods can detect thousands of compounds, but there’s a problem: current technology can only identify about 10% of them. This project aims to change that.

By using probabilistic and transformer deep learning models, we will analyze complex water samples from lakes, rivers, drinking water, and wastewater faster and more accurately than ever before, spotting hidden chemical fingerprints in water samples and predicting what those unknown compounds might be.

With climate change and pollution threatening global water supplies, this technology could redefine how we safeguard water quality, making sure the environment we rely on is safe and clean for future generations.