Technical University of Sofia | Persy Case Study

Technical University of Sofia | Persy Case Study

NVIDIA H200. 100GbE. AI infrastructure for the next generation of engineers.

2 GPU servers. 12 AI workstations. A platform built with Persy.

Persy delivered a GPU-accelerated computing platform for the Technical University of Sofia, designed for neural network training, scientific computing and advanced AI development. The project combines NVIDIA H200 NVL accelerators, high-speed 100GbE networking and a modern architecture optimized for intensive GPU workloads.

The solution creates an infrastructure that connects academic education with real enterprise technology.

Key Metrics

GPU Infrastructure

  • 2 ESC8000A-E13 GPU server systems
  • NVIDIA H200 NVL accelerators
  • AMD EPYC processors and DDR5 memory
  • 100GbE high-speed connectivity

AI Workstations

12 ASUS GX10 Compact Desktop AI Supercomputer systems

Project

Value: ≈ 114,274 EUR excl. VAT

Delivery, deployment and training

Context

Modern engineering and research programs require access to real AI infrastructure. The university needed a platform enabling AI model training, large-scale data processing and hands-on experience with enterprise GPU technologies.

The Challenge

The solution had to provide:

  • high compute power for neural network training;
  • stable performance under intensive GPU workloads;
  • high-speed data exchange;
  • scalability for future infrastructure growth.

Persy’s Engineering Approach

Persy designed an architecture combining centralized GPU compute power with individual AI workstations. ESC8000A-E13 GPU servers with NVIDIA H200 NVL accelerators, PCIe 5.0 architecture and 100GbE networking were deployed to support advanced AI training and scientific simulations.

All activities — delivery, installation, configuration, testing and staff training — were executed on schedule. 

The Result

Following deployment, the Technical University of Sofia now operates a modern AI platform that:

  • accelerates neural network training;
  • supports scientific research;
  • provides hands-on access to enterprise GPU infrastructure;
  • enables future AI development.

Strategic Impact

The project demonstrates how modern GPU infrastructure transforms engineering education. By implementing NVIDIA H200-based architecture, Persy contributes to building an environment for the next generation of AI specialists and research innovation.

View all →