Cornell Tech is bringing on six new faculty members whose research is already influencing how artificial intelligence systems learn, reason, and solve real-world problems. These scholars have created programming languages for high-performance computing, developed widely-used benchmarks for machine unlearning, and built AI systems that learn more efficiently from experience. Their collective work addresses some of the field's most pressing challenges: making AI more trustworthy, adaptable, and capable of accelerating scientific discovery.

What Are These Researchers Actually Working On?

The six new faculty members bring expertise spanning artificial intelligence, machine learning, programming languages, and operations research. Their accomplishments have already earned recognition from leading conferences and competitive fellowships, including the Quad Fellowship, OpenAI Research Award, and Association for Computational Linguistics Outstanding Paper Award.

One faculty member, Yuka Ikarashi, created the Exo programming language, which enables developers to write high-performance code on hardware accelerators. This work addresses a critical bottleneck in AI development: making it easier to deploy models efficiently on specialized chips. Ikarashi has received multiple prestigious awards and previously worked at Apple, Amazon, and CERN, applying her research to a range of accelerators and applications.

Another researcher, Vedavyas Setlur, focuses on building AI systems that can continually adapt and improve during use, including foundation models that dynamically scale computing resources for reasoning and exploration as they solve complex problems. His work has been published in leading machine learning venues including ICLR, ICML, NeurIPS, and AISTATS, with multiple spotlight and oral presentation selections.

Vivek Maini studies how the data used to train AI systems shapes what they learn, remember, and how reliably they behave. His research focuses on data-centric AI, including improving pretraining data, generating synthetic data, and understanding unwanted memorization in foundation models. His benchmarks for machine unlearning have become widely used throughout the field, and his work on synthetic data and AI safety has informed practices at leading AI labs.

How Are These Advances Changing AI Development?

• Data-Centric Approaches: Researchers like Maini are shifting focus from simply building larger models to carefully curating and understanding the data that trains them, which improves reliability and reduces harmful memorization in AI systems.
• Efficient Reasoning: Faculty members are developing AI systems that can dynamically allocate computing resources based on problem complexity, making AI more efficient and practical for real-world deployment.
• Learning from Experience: New research in reinforcement learning and interactive learning enables AI systems to adapt to new environments and acquire new capabilities more efficiently, rather than requiring complete retraining.
• Trustworthy AI: Work on machine unlearning and AI safety is creating methods to help AI systems become more controllable, factual, and transparent about their limitations.

Abhishek Sekhari, currently a senior research scientist at the Chan Zuckerberg Biohub, focuses on reinforcement learning and interactive learning with the goal of developing machine learning systems that learn efficiently from experience, adapt to new environments, and acquire new capabilities. His work spans reinforcement learning, optimization, machine unlearning and privacy, and AI for science, drawing on both theoretical and empirical methods. He previously held a postdoctoral position at MIT and earned his Ph.D. from Cornell University.

Weijia Shi develops augmented and modular architectures and training algorithms that make language models more controllable, collaborative, and factual. Her work has advanced methods for helping AI systems use external tools and specialized models, improving their ability to access information, identify knowledge gaps, and reduce hallucinations. Her techniques have been adopted by researchers and integrated into platforms such as Databricks MosaicML, LlamaIndex, and LangChain. She received an Outstanding Paper Award at ACL 2024 and was named a Rising Star in Machine Learning in 2023 and in Data Science in 2024.

Rad Ma's research focuses on decision-making under uncertainty, online algorithms, and AI agents for operations research. He is recognized for contributions to revenue management, online matching, inventory management, learning theory, e-commerce fulfillment, and neural network verification. Algorithms from his research have been deployed at companies including Alibaba, Bed Bath & Beyond, and Dream11. Before joining Cornell Tech, Ma was the Roderick H. Cushman Associate Professor at Columbia Business School and earned his Ph.D. from MIT.

"We're thrilled to welcome this exceptional cohort of faculty to Cornell Tech. These scholars have already established themselves as leaders in fields ranging from machine learning and AI safety to programming systems and operations research, producing work that is influencing both academic research and industry practice," said Greg Morrisett, Jack and Rilla Neafsey Dean and Vice Provost of Cornell Tech.

Greg Morrisett, Jack and Rilla Neafsey Dean and Vice Provost of Cornell Tech

Why Does This Matter for the Future of AI?

These faculty additions represent a strategic investment in addressing fundamental challenges that will define AI development for years to come. Rather than simply scaling up model size, this cohort is tackling questions about efficiency, trustworthiness, and how AI can be made more adaptable and responsible. Their work spans both theoretical advances and practical applications that are already being deployed in industry.

The research areas represented by these six faculty members reflect a broader shift in AI research away from one-size-fits-all approaches toward more specialized, efficient, and interpretable systems. By focusing on data quality, learning efficiency, and AI safety, these researchers are helping build a foundation for AI that is more capable, reliable, and impactful across science, industry, and society. Their arrival at Cornell Tech deepens the institution's ability to tackle the technological challenges that will define the coming decades.