[HPC Case Study] In-house Local LLM Development and Construction Use HPC

We are utilizing it as an AI infrastructure that can be gradually expanded from PoC to full-scale operation!

We would like to introduce a case study on the implementation of "HPC" for the development and construction of in-house local LLMs. To safely utilize highly confidential information such as design documents, analysis data, quality reports, and research materials, we built a local LLM infrastructure in an on-premises environment due to concerns about information leakage and external data storage with cloud-based generative AI. By introducing GPU-equipped HPC servers, we aimed to achieve a secure generative AI operation that is self-contained within the company's closed environment. Furthermore, with high-speed inference processing using high-performance GPUs, we can smoothly execute tasks such as searching and summarizing large documents, responding to technical inquiries, and utilizing design knowledge, thereby enabling the generation of high-precision responses based on internal documents and expecting improvements in operational efficiency for the technical and research and development departments. 【Workstation Specifications (Partial)】 ■ Basic System: Windows 11 Pro 64bit ■ Chassis: Mid-tower chassis (W) 235 × (D) 470 × (H) 495 mm ■ Power Cable: 100V power cable *For more details, please refer to the related links or feel free to contact us.

Industrial PCs