DuckAI 2025 Fall

In this talk, Dr. Tianhao Wu, founder and CEO of opAIda, presents a practical and scalable strategy for unlocking business ROI through a Data All-in-One architecture powered by open-source AI. The approach emphasizes affordability, data ownership, and operational control by combining open-source LLMs with low-cost, high-performance hardware.

Dr. Wu introduces the Unified Data Agent (UDA)—a customizable AI agent designed to eliminate the pervasive challenge of enterprise data silos. UDA unifies disconnected data sources such as email, chat, cloud drives, and databases without requiring any manual work. By learning directly from real human actions and institutional processes, the system automates complex workflows including RFI responses, event operations, task prioritization and cross-team coordination. The result is a continuously improving AI ecosystem where organizational knowledge is preserved, searchable, and actionable through semantic retrieval. This Data All-in-One framework demonstrates how open-source AI can deliver immediate efficiency gains while laying the foundation for long-term digital transformation.

Short Bio: Dr. Tianhao Wu is a trailblazer in artificial intelligence and a visionary leader with over 24 years of experience transforming complex business challenges into scalable AI solutions. Before co-founding opAida.ai, he founded AYR, Inc., where he pioneered the Singularity AI platform. In 2023, following a rigorous two-year evaluation of more than 50 vendors—including Microsoft and Google—his AI technology was selected exclusively by the U.S. Internal Revenue Service to automate all paper-filed tax returns nationwide. A hands-on innovator, Dr. Wu has delivered over 200 machine learning use cases across government and highly regulated industries, with solutions powering platforms for the U.S. National Institute of Justice, Lockheed Martin, and WorkFusion. He holds a Ph.D. in Machine Learning from Lehigh University and a B.Eng. in Computer Science from Peking University, with four granted patents and three pending. His early work building AI-powered search technology at Ask.com further solidifies his reputation as a pioneering force in enterprise AI.

Dr. Yingchao Zhang, widely known as “Dr. YZ,” is a highly respected technology executive with over two decades of experience spanning system administration, cybersecurity, software engineering, and AI innovation. A nuclear physicist by training, he made a bold pivot into tech in 1997 and has since led transformative initiatives across startups, research institutions, and enterprise environments—including key roles at NetScout serving Fortune 500 clients. He holds a Ph.D. in Nuclear Physics from Stony Brook University and a B.Sc. from the University of Science and Technology of China. In recognition of his leadership, Dr. YZ was named one of the Outstanding 50 Asian Americans in Business in 2023 and appointed in 2025 by New Jersey Governor Phil Murphy to the Board of Trustees of Thomas Edison State University. At opAida.ai, he brings strategic vision, deep technical expertise, and operational excellence to the company’s mission of delivering secure, scalable AI solutions.

Large Language Models (LLMs) have been applied to expert domains and scientific contexts, such as clinical diagnosis and drug discovery. However, the generalizability that characterizes LLMs in the general domain does not readily translate to scientific and expert tasks. Unlike general natural language tasks, scientific data is densely packed, homogeneous, and less self-explanatory. Moreover, expert-level tasks, such as those performed by physicians, engineers, or scientists, require deep domain knowledge, intuitive reasoning, and multi-step planning, refined through years of specialized training. In this talk, I will first discuss capturing implicit expert intuition for individual decision-making, using clinical diagnosis prediction as a case study. I will then extend the focus to compositional, project-level reasoning and automation, highlighting the development of LLM agents for scientific discovery.

Short Bio: Mingyu Derek Ma is a Senior Machine Learning Scientist at Prescient Design, Genentech (Roche), where he leads the development of agentic intelligence and automation platforms for drug discovery and contributes to training scientific large language models. His work focuses on the architecture, training, and agentic use of generative language models inspired by and applied to clinical, medical, and scientific scenarios. His research has been recognized as one of the top 15 most-cited papers at NAACL 2024 and has been published at leading AI conferences like ACL, AAAI, and NeurIPS. He obtained his PhD in Computer Science from UCLA and received the J.P. Morgan Chase AI PhD Fellowship and Amazon Fellowship.

Ensuring the security and safety of autonomous driving systems is a critical challenge due to the inherent safety risks. Building safe and human-aligned autonomous driving systems in real-world scenarios remains unresolved due to the complex and varied nature of driving behaviors and semantics. In this talk, I will present recent research efforts aimed at rethinking the testing of AV systems. The discussion will focus on two key areas: 1) modeling diverse behaviors using advanced traffic models, and 2) leveraging foundation models to enhance simulation capabilities and accelerate the testing process.

Short Bio: Yulong Cao is a research scientist at the NVIDIA Autonomous Vehicle Research group directed by Marco Pavone. His research interest is in the intersection of security, privacy, and machine learning. His research goal is to build trustworthy machine learning systems that are aligned with human in the real world, with a focus on autonomous driving systems. He completed his Ph.D from the Computer Science and Engineering department of University of Michigan, advised by Morley Mao. and holds B.S. degrees from University of Michigan and Shanghai Jiao Tong University.