I am Javier Rando, a Doctoral Student at ETH Zurich advised by Florian Tramèr and Mrinmaya Sachan. My research tries to answer the question "What will go wrong when we deploy powerful AI models in real-world applications?" and usually involves red-teaming frontier LLMs. Some of the research directions I am currently thinking about are (1) scalable red-teaming, (2) LLM poisoning, (3) detecting emergining hazardous capabitilies, (4) security risks of LLMs (as agents) and their implications. Before starting my doctorate, I obtained a Computer Science MSc from ETH Zurich and a Data Science BSc from Pompeu Fabra University. I also was a visiting researcher at NYU under the supervision of He He and founded EXPAI, an explainable AI startup in Spain.