Architecting Physical Intelligence: Cross-Stack Co-Design from Systems to Silicon
11:40 AM to 1:00 PM
CSB 451 CS Auditorium
Zishen Wan, Harvard University
Abstract:
Physical intelligence – where embodied agents perceive, reason, plan, and act in the physical world – is emerging as a new computing frontier spanning robotics, autonomous systems, and spatial AI. However, today’s physical intelligence systems remain constrained by high latency, energy cost, and fragile reliability, due to fundamental mismatch between their compositional nature and existing computing architectures. The core challenge extends beyond algorithms, to how we architect computing systems and silicon that natively support intelligence that reasons and adapts under real-world constraints.
In this talk, I will present a principled cross-stack system-architecture-silicon co-design approach to building the computational foundations for physical intelligence. (1) First, I will introduce CogSys, a flexible hardware architecture for efficient neuro-symbolic cognition. CogSys features unified kernel abstractions and reconfigurable dataflows, delivering notable efficiency improvements across neuro-symbolic models and validated on FPGAs. (2) Next, I will showcase the first programmable SoC tapeout for neuro-symbolic cognition, demonstrating how tightly integrated memory-centric computing, heterogeneous architectures, end-to-end compilation flow, and adaptive power management enable efficient cognition in silicon. (3) Building on this foundation, I will present ReCA, an integrated hardware architecture that bridges high-level cognition and low-level autonomy under stringent power and latency constraints by leveraging spatial-aware runtimes, heterogeneous fabrics, and hybrid memory hierarchies. (4) Finally, I will highlight our agile SoC design flows that translate evolving cognition and autonomy workloads into efficient silicon implementations.
By bridging computer architecture, system software, and silicon validation, my research establishes adaptive, accelerator-rich computing substrates for physical intelligence. This work advances a vision in which AI and hardware are co-designed, co-reason, and co-adapt, architecting future computing systems as active enablers of intelligence in the physical world.
Add to GMail
Scaling RL Rollouts: Agent-Native Infrastructure with Daytona
11:30 AM to 1:00 PM
Davis Auditorium
Ivan Burazin
Abstract:
In this talk, we’ll outline why a new class of agent-native infrastructure is emerging, what problems it is designed to solve, and the core use cases driving it, from autonomous coding agents to large-scale evaluation and training workloads. Daytona is an agent-native control plane designed to orchestrate isolated, stateful sandbox environments at scale. We’ll break down the infrastructure challenges behind isolation, state management, and massive parallelism, and why traditional VM and container stacks fall short. As a concrete example, we’ll walk through scaling RL rollouts, showing how tens of thousands of environments can be provisioned and orchestrated in minutes as part of a high-throughput RL pipeline.
This event is organized by Columbia's Data, Agents, and Processes Lab (DAPLab). For more information about the series, see https://daplab.cs.columbia.edu/entrepreneurship.
The Columbia Engineering AI Entrepreneurship Series is a bi-weekly speaker series that brings students and faculty at Columbia together with founders, VCs, technologists, and business leaders to learn about the process of transitioning lessons from research and the classroom into products and value.
Hack-a-Technical Interview with Visa
12:30 PM to 1:30 PM
CS Lounge
Visa will host an exclusive workshop to provide students with valuable insight into its recruiting and application process. This session is a great opportunity to gain a competitive edge and better understand what it takes to stand out as a candidate. During the workshop, the Visa team will walk through the application process in detail, share insider strategies on how to differentiate yourself, and provide an in-depth breakdown of interview preparation. Recruiters will also offer guidance on strengthening your resume and effectively positioning your experiences. Don’t miss this opportunity to connect directly with the recruiting team, ask questions, and gain clarity on how to navigate future recruiting cycles with confidence.
Registration Information will be posted via email, VMock and CampusGroups.
*Event Audience: CS Graduate Students (MS/PhD) & Bridge Students
Blackrock Coffee Chat Session
11:00 AM to 12:00 PM
Zoom
Join us for an exclusive coffee chat with Shi Tak (Dave) Man, BlackRock senior engineering leader who brings over 17 years of experience delivering large-scale initiatives from concept through launch across global teams. With 14 years of people management experience, they offer extensive expertise in leadership, team development, and project execution. This is an excellent opportunity for students to gain insights into engineering career paths, effective leadership, and leading impactful, cross-functional work.
If you would like to be considered for a small group coffee chat session with Shi Tak (Dave), please complete the Google interest form. The form has been distributed via email and VMock. Deadline to complete the form is Monday, February 23rd at 4 PM ET.
*Event Audience: CS Graduate Students (MS/PhD) & Bridge Students.
Foundations of Language Models: Scaling and Reasoning
11:40 AM to 12:40 PM
CSB 451 CS Auditorium
Eshaan Nichani, Princeton University
Abstract:
Modern deep learning methods, most prominently language models, have achieved tremendous empirical success, yet a theoretical understanding of how neural networks learn from data remains incomplete. While reasoning directly about these approaches is often intractable, formalizing core empirical phenomena through minimal “sandbox” tasks offers a promising path toward principled theory. In this talk, I demonstrate how proving end-to-end learning guarantees for such tasks can provide a practical understanding of how the network architecture, optimization algorithm, and data distribution jointly give rise to key behaviors. First, I will show how neural scaling laws arise from the dynamics of stochastic gradient descent in shallow neural networks. Next, I will study how and under what conditions transformers trained via gradient descent can learn different reasoning primitives, including in-context learning and multi-step reasoning. Altogether, this approach builds theories which yield concrete insight into the behavior of modern AI systems.
Add to GMail