Rafael Gómez-Bombarelli, an MIT associate professor in materials science and engineering, has spent over a decade harnessing artificial intelligence to accelerate scientific discovery. Now, with the rapid evolution of AI, he sees a transformative moment on the horizon—one where AI, simulations, and multi-modal models converge to create what he calls “general scientific intelligence.” Gómez-Bombarelli, who was recently tenured, believes the field has reached a second major inflection point. The first came around 2015, when representation learning and early generative AI began reshaping scientific research. His lab at MIT was among the pioneers, applying machine learning and deep learning to molecular modeling and materials design. Today, he envisions a future where AI can reason across languages, material structures, and synthesis pathways—integrating diverse forms of scientific knowledge into a unified framework. His work blends physics-based simulations with AI to discover new materials with real-world impact. His team has developed novel compounds for batteries, catalysts, plastics, and organic light-emitting diodes (OLEDs). Beyond academia, Gómez-Bombarelli has co-founded several startups, including Lila Sciences, a company building a scientific AI platform for life sciences, chemistry, and materials research. His goal is to make scientific discovery faster, more efficient, and more accessible. Born in Spain, Gómez-Bombarelli was drawn to chemistry early, winning a national chemistry competition in 2001. He earned his undergraduate degree at the University of Salamanca and later completed a PhD focused on DNA-damaging chemicals. Midway through his doctorate, he became fascinated by computational methods, shifting from lab experiments to simulations. He found programming to be a powerful way to structure complex scientific thinking—less constrained by physical tools than traditional experimentation. His career took a pivotal turn during a postdoc in Scotland, where he connected with Alán Aspuru-Guzik at Harvard. There, Gómez-Bombarelli became one of the first researchers to use generative AI for molecular design in 2016 and helped pioneer neural networks for chemistry in 2015. He began automating simulations to run hundreds of thousands of calculations, uncovering promising materials for further testing. After two years, he and Aspuru-Guzik launched a materials computation company that eventually focused on OLEDs. Running it full-time was challenging but rewarding, teaching him the value of turning research into tangible products. However, seeing Aspuru-Guzik’s demanding schedule made him hesitant about an academic career—until his mentor insisted he apply for a position at MIT. At MIT, Gómez-Bombarelli found a research environment that matched his vision. He now leads a fully computational lab of about 25 graduate students and postdocs, focusing on how atomic composition and structure influence material behavior. His team develops tools that combine deep learning with physics-based models, creating a feedback loop where more data improves both AI and simulation accuracy. Though they don’t conduct physical experiments, his group collaborates closely with experimentalists, helping them prioritize the most promising AI-generated ideas. He also works with industry partners through MIT’s Industrial Liaison Program to ensure research addresses real-world challenges. As AI has matured, Gómez-Bombarelli has seen widespread adoption of his early approaches. Companies like Meta, Microsoft, and Google DeepMind now use AI-driven simulations, and the U.S. Department of Energy’s Genesis Mission underscores the national push to leverage AI for scientific advancement. He believes we’re at a turning point where large language models—mastered in natural language—can now accelerate science by bridging the gap between human thought, written research, and computational insight. “We’ve seen scaling work for language and for simulations,” he says. “Now we’re seeing how it works for science.” Gómez-Bombarelli carries forward the collaborative, non-competitive spirit he found at MIT. He now encourages his students to pursue faculty positions—even if they miss the deadline—passing on the mentorship he once received. For him, the future of science isn’t just faster discovery—it’s a more inclusive, intelligent, and human-centered journey.