BDI High Performance Workstation
The Digital Scholarship Lab has a High Performance Workstation available for student projects and experimentation. This resource is not part of the Brandeis High Performance Computing Cluster.
BDI's High Performance Workstation has an Intel(R) Xeon w7, 256 GB RAM, 196 VRAM, +5 TB Drive RTX 6000. It has both Linux and Windows operating systems and can be accessed in the DS Lab, or via Teamviewer.
If you require a Linux operating system but don't need this level of hardware, please reach out to makerlab@brandeis.edu. We can help you build a bootable drive.
Expand All
If you have questions about whether this workstation is appropriate for your project, or you are interested in learning more about LLMs, image recognition, and other applications, please reach out to
Tim Hebert.
Natalie Susmann can help with big-data GIS, and photogrammetry projects, and Humanities and Social Sciences questions.
Lillian Yin '26, Business and Economics major
Lillian is using simulation to analyze the effect of algorithmic traders on market microstructure in equity markets; requiring high-performance computing to train agents in deep learning models on L2 historical data to validate simulation results and assess whether they hold in real markets.
Alex Ott '27, Computer Science
Alex used the system to run local LLMs via LM Studio, exposing the models both through the LM Studio desktop interface and remotely through OpenWebUI on the university network. Most experimentation focused on running gpt-oss-120b, which is impractical to run efficiently on typical consumer hardware. Using LM Studio’s OpenAI-compatible API, Alex built custom tooling that integrated the local models into AI-assisted software development workflows, enabling tasks such as summarization, information extraction, and revision over sensitive material without sending data to third-party services. These experiments informed the development of a workshop and written guide introducing LLM fundamentals and practical, privacy-preserving and academic use cases. In addition, Alex collaborated with another student who used the same system to develop and train a stock prediction model, gaining hands-on experience with the constraints and workflow considerations of developing on a multi-GPU machine.
During his tenure as a BDI student worker, Alex helped with the acquisition and maintenance of the HPW, and also sparked new collaborations between the Brandeis Hack Club and BDI via the HPW.