Machine Learning for Chemistry: Accelerating Discovery with Intelligence
In our lab, we are redefining how chemistry is done, not by replacing human intuition, but by expanding it. We use machine learning to uncover patterns in molecular data that are often invisible to the human eye. Through contrastive learning, graph-based models, and fingerprint embeddings, we’re building systems that don’t just predict outcomes but help us understand why reactions behave the way they do. Instead of relying on trial-and-error, we aim for insight-driven design — algorithms that learn from reactions, refine their own understanding, and evolve with every dataset. We train models to grasp reactivity, stability, and selectivity at a fundamental level, enabling faster and more intelligent decision-making in catalysis and synthesis. Our goal is to turn molecules into data, and data into discovery — forging a new kind of chemistry that is faster, cleaner, and far more curious.
Drug Discovery with AI: Imagining Medicines Before They Exist
We believe the future of drug discovery lies not just in automation, but in imagination, and that machines can help us dream better. In our group, artificial intelligence is a creative partner that generates new bioactive compounds, predicts activity, and learns from every success and failure. Using contrastive learning, transformer models, and generative frameworks, we explore vast chemical spaces and propose molecules no one has seen before. These models are trained to think in molecular language. We integrate docking scores, biological data, and molecular structure into a unified pipeline that designs and refines drug candidates with minimal supervision. Our vision is to build systems that accelerate discovery while keeping human insight at the center. With every new molecule imagined by a machine, we get one step closer to medicines that are faster to find, easier to make, and more precisely targeted than ever before.
