Cal Poly opens $3 million AI Factory for student and faculty research

Direct access to high-end hardware

Cal Poly has officially unveiled a new AI Factory designed to give students and faculty direct access to advanced computing resources. The facility was funded by the Noyce School of Applied Computing and features four NVIDIA DGX B200 systems. These systems are paired with the necessary storage and networking infrastructure required to train large AI models on university-controlled resources.

The school launched the facility during an event called AI Day on October 22. By hosting this hardware on campus, the university aims to provide a space where students can work with systems that are too large for a typical laptop to handle. This move is intended to help graduates understand the limits of AI so they can better evaluate its outputs in the workforce.

Computer science professor Alex Bisberg noted that owning the hardware provides the school with greater control over how students learn. He explained that while many students can run small models on personal computers, the scale of modern datasets requires much more power. By providing this infrastructure, the school ensures that students can interact with the same level of technology used by large private companies.

Privacy and local model training

One of the primary advantages of an on-campus AI Factory is the ability to process sensitive information securely. Because protected patient information cannot be shared with external AI services, researchers need a way to train models locally. The new hardware allows for the analysis of complex medical data without risking privacy violations.

Professor Paul Anderson is currently using the facility to work on projects involving spinal cord injuries and genetic risks. He noted that before the AI Factory was established, his team could not work on these specific problems because they lacked the necessary compute. Now, the university has the power to train frontier models on these types of high-stakes datasets.

This capability changes how the university can approach public health and medical research. Instead of relying on third-party APIs that might pose security risks, faculty can keep all data within the university's walls. This allows for a more robust exploration of patterns in complex medical data while maintaining strict security protocols.

Accelerating environmental research

The AI Factory is already being used to improve the speed of environmental monitoring and data analysis. Professor Jonathan Ventura is applying machine learning to drone imagery of eelgrass in Morro Bay to study the local ecosystem. Previously, researchers had to annotate these images by hand, which was a slow and labor-intensive process.

The new system matched the quality of the manual analysis but processed the images roughly 10 times faster. This increase in speed allows for more efficient monitoring of marine environments. It demonstrates how high-performance computing can turn a bottleneck into a streamlined research workflow for environmental scientists.

By automating the analysis of drone footage, the university can produce results more quickly than human observation alone. This type of application shows how the AI Factory serves practical needs beyond just theoretical computer science. It provides a tool for researchers to manage large-scale visual data that would otherwise be impossible to process in a reasonable timeframe.

Student-led tools and ethics

Students at Cal Poly are also using the new resources to develop practical tools that address the ethical and functional aspects of AI. For example, student Jillian Bogosian is creating a tool for legal professionals to help them assess if digital images submitted as evidence have been altered. Her project involved working directly with lawyers to ensure the tool provides clear indicators for human judgment.

Another student project, Orbit.IQ, is a browser tool designed to summarize a student's AI activity. It includes a prompt budget intended to encourage users to make more deliberate choices when interacting with AI. This project grew from a team effort to help peers use AI tools while still preserving the critical thinking skills required for learning.

These projects highlight the university's goal of teaching students how to contribute meaningfully to the development of AI. By training the models themselves, students can learn about the underlying systems and research new ways to improve them. The school aims to prepare students to make informed technical and ethical choices as they enter the professional world.

The role of the Noyce School

The AI Factory is a key component of the Noyce School of Applied Computing, which was established through a $60 million gift from the Robert N. Noyce Trust. This school is the first of its kind at Cal Poly and is designed to be interdisciplinary. It brings together several fields, including Electrical Engineering, Computer Science, Software Engineering, and Computer Engineering.

The school is housed within the College of Engineering and uses Statistics as an affiliate. This structure is designed to create opportunities for students and faculty across different computing fields to collaborate. By working together, they can solve complex problems using shared computational principles and technologies.

The establishment of the school provides the administrative and academic framework for large-scale projects like the AI Factory. It ensures that the hardware is integrated into a broader educational mission. The goal is to provide a holistic education where students learn the tools of AI while also understanding the mechanics that power them.