Victor Petre

Background

About Victor Petre

A fan of happy accidents, unexpected discoveries, and following curiosity wherever it leads. My academic journey has been defined by a pursuit of what’s interesting over what’s straightforward. I began with a broad fascination with natural sciences, without a single thought of working in artificial intelligence. That changed in my final undergraduate year when I encountered computational approaches to modeling cognition, a moment that sent me down an entirely new career path. Over the years I acquired a broad range of experiences, from Kaggle competitions and generative modeling to medical imaging, text mining, and neuromorphic computing. Although some of my more obscure projects didn’t quite hit their intended mark (turns out that “particle-based slime mold simulations” aren’t the ideal approach to railway design, and dragonfly-inspired algorithms may not be great at classifying EEG data), each project has added to a skill set I keep finding new ways to apply. What excites me most is how simple, elegant mathematical ideas can drive major breakthroughs in AI. Currently, I’m researching efficient training methods for deep-learning-based sequence models, with the intent of moving beyond analytical gradient computation altogether. I’m always open to connecting with others driven by curiosity and innovation. Shoot me a message if you’re interested in exploring new ideas in AI, data science, or beyond!