Ce Zhang is a computer scientist and academic known for contributions to machine learning systems and data management. He is an associate professor at the University of Chicago and a co-founder of Together AI, a company focused on open-source generative AI infrastructure. His research bridges Artificial intelligence with database and systems design, aiming to make Machine learning more efficient and accessible.
Zhang's work has been recognized with multiple awards, including a Sloan Research Fellowship and an NSF CAREER Award. He has published extensively in top venues such as SIGMOD, VLDB, and NeurIPS, and his research has influenced both academic and industrial practices in deploying large-scale AI models.
Early Career and Education
Zhang received his Ph.D. in computer science from the University of Toronto, where he worked under the supervision of Renée J. Miller. His doctoral research focused on data integration and cleaning, which later informed his interest in applying these techniques to machine learning pipelines. Before joining the University of Chicago, he was a postdoctoral researcher at Stanford AI Lab, collaborating with Christopher Ré on the DeepDive system, which used statistical inference for information extraction from unstructured data.
Research Contributions
Zhang's research centers on the intersection of data management and machine learning. He has developed systems that optimize the training and inference of Deep learning models by leveraging database principles such as indexing, compression, and query optimization. One notable project is the "Hogwild!"-style parallel SGD implementation, which demonstrated how to scale Stochastic Gradient Descent Variants on multicore systems without locking, achieving near-linear speedups.
He also contributed to the development of "Cerebro," a system for large-scale model selection and management, which allows data scientists to efficiently explore hyperparameter spaces. His work on "LazyBatching" and "DataSkipping" techniques has improved the performance of Neural network training on relational data, reducing I/O and computation costs. These contributions have been widely cited and have influenced subsequent research in the field of ML systems.
Together AI and Industry Impact
In 2022, Zhang co-founded Together AI alongside other researchers and engineers. The company provides cloud infrastructure for training and running Large language models, with a focus on open-source models and tools. Together AI's platform supports popular models like Llama and Mistral, offering services for fine-tuning, inference, and deployment. The company has raised significant funding and has become a key player in the Generative AI ecosystem, competing with larger cloud providers by offering specialized, cost-effective solutions.
Zhang's role at Together AI involves guiding the technical direction, particularly in optimizing the performance of Transformer (architecture) architectures on distributed systems. His academic insights have directly shaped the company's approach to building scalable AI infrastructure, and he continues to bridge research and industry practice.
Teaching and Mentorship
At the University of Chicago, Zhang teaches courses on machine learning systems and data management. He has mentored numerous Ph.D. students and postdocs, many of whom have gone on to positions in academia and industry. His teaching emphasizes hands-on experience with building and deploying AI systems, encouraging students to understand both theoretical foundations and practical engineering challenges.
Zhang has also been active in organizing academic conferences and workshops, serving on program committees for major venues like NeurIPS, ICML, and SIGMOD. He is a strong advocate for reproducibility in AI research, and his lab maintains open-source code and benchmarks for the systems they develop.
Recognition and Awards
Zhang's contributions have been recognized with several honors. He received the Sloan Research Fellowship in 2021, which is awarded to early-career scientists of outstanding promise. He also received the NSF CAREER Award, a prestigious grant for junior faculty. His paper on "Cerebro" won the Best Paper Award at ICDE 2020, and his work on "Hogwild!"-style algorithms has been recognized as a foundational contribution to parallel machine learning.
He has been invited to give keynote talks at various conferences and workshops, and his research has been featured in media outlets covering AI advancements. His work with Together AI has also been highlighted in industry reports on the growing landscape of AI infrastructure providers.
Future Directions
Zhang continues to explore new frontiers in machine learning systems, including efficient training methods for Residual Network (ResNet)s and other architectures, as well as techniques for Model Pruning and Data Augmentation. He is interested in making AI more sustainable by reducing the computational resources required for training large models. His ongoing research aims to develop systems that can adapt to changing data distributions and support real-time inference at scale.
As a co-founder of Together AI, Zhang is also focused on democratizing access to AI, ensuring that researchers and developers can leverage powerful models without prohibitive costs. His dual role as an academic and entrepreneur positions him to influence both the theoretical and practical evolution of artificial intelligence in the coming years.