About Me

Calder Katyal is a passionate student of Applied Mathematics and Computer Science at Yale University who concentrates in Machine Learning. With the ability to learn extremely quickly and adapt to new learning environments, he is constantly exploring new technologies. Calder has researched and published in various fields of ML, such as geometric learning, convex optimization, and reinforcement learning; he has also interned at AI startups doing advanced technical research. In Summer 2025, Calder worked as a research intern on the multimodal (video generation) team at Character.ai, a Silicon Valley startup founded by Noam Shazeer. He successfully developed a complete end-to-end data preprocessing pipeline using Apache Spark, hundreds of H100 GPUs, and open-source models to curate a dataset of millions of videos and hundreds of thousands of hours of footage—accelerating preprocessing speed by 4x. He furthermore collaborated with former ByteDance researcher Weimin Wang to train and distill a 13B-parameter, Stable Diffusion–style model that generates high-quality, audio-synced videos. In Summer 2024, Calder worked as a Machine Learning Intern at the Palo Alto-based startup CloudChef, Inc., where he leveraged advanced computer vision techniques and frameworks, synthesizing a massive dataset with thousands of multimodal data points and training a high-performing embedding model for cooking footage despite limited compute and noisy data. He has previously worked for Appian Corp, a multibillion cloud-computing and enterprise software company, reporting directly to the founder. Calder is experienced in Python, PyTorch, C/C++ and more, taking graduate-level ML courses at Yale. Most importantly, Calder is a well-rounded student, passionate about all things from math to journalism (published author on The Atlantic). In his free time, you can find Calder reading, playing classical piano, or singing.