CoDEx 2026 Sponsor Presentations
Thank you to our CoDEx 2026 corporate sponsors for providing the following presentations, which will be held in the Arch Room.
2:05 to 2:25 p.m. - Comsol
Accelerating Simulation Apps with Data - Driven Surrogate Models
Akhilesh Sasankan, Senior Application Engineer, COMSOL, Inc.
High-fidelity simulations are essential for accuracy but can be computationally expensive, often limiting real-time analysis and rapid design iteration. This session will demonstrate how to overcome these limitations by using data-driven surrogate models, such as deep neural networks (DNN), to create fast, lightweight approximations of full finite element models.
We will discuss the workflow in the COMSOL Multiphysics® software for generating training data using efficient geometry sampling, as well as training these surrogates to achieve near-instantaneous results. Through live demonstrations of simulation apps, you can get a better idea of how surrogate models enable seamless user experiences and wider adoption of simulation across organizations.
2:35 to 2:55 p.m. - Cambridge Computer and Nvidia
Accelerating Higher Education Science and Research with Cambridge Computer and NVIDIA
Jose Alvarez, Vice President, Research Computing HPC/AI, Cambridge Computer
Cambridge Computer, a multi-time NVIDIA Higher Education Partner of the Year, helps universities and research institutes design, procure, and operate GPU-accelerated HPC and AI infrastructures that align with real scientific and budget realities.
By combining NVIDIA’s full-stack portfolio—DGX systems, data center GPUs, networking, and software platforms for AI, data science, and simulation—with Cambridge’s PhD-level research computing team, institutions can rapidly stand up or expand AI clusters, shared research facilities, and cloud-hybrid environments that power teaching, cross-campus collaboration, and breakthrough discovery across disciplines from life sciences to climate modeling.
3:05 to 3:25 p.m. - Google Cloud
Scientific Literature Assistants: Embedding Spaces and Retrieval Augmented Generation for AI Applications
John Cecala - Google Higher Education Advocate | Research & AI | Student Success
- For half a century, one of the grand challenges in medicine was predicting how proteins fold into their complex 3D shapes. A problem so difficult it was thought to be unsolvable. Google AI solved it.
- In Materials Science, discovering a new stable material, such as those used in batteries or solar panels, can take decades. Google AI analyzed over 380K hypothetical materials, identifying over 1,000 new stable compounds, effectively condensing 800 years of research into a matter of weeks.
- Some mathematical algorithms used in computer science have been the same for over 50 years. Google AI discovered a more efficient way to perform matrix multiplication that no human mathematician had identified.
