Cloud4Y provided GPU infrastructure to Vitasoft

2024: Infrastructure provision

Cloud4Y provided infrastructure with GPU for JSC Vitasoft. This was announced by Cloud4Y on November 6, 2024.

The company uses cloud infrastructure for machine learning projects.

Since 2005, Vitasoft JSC has been operating on the market, and since 2010 it has been included in the register of accredited IT companies. Vitasoft has competencies and resources for working with hardware and software complexes of world manufacturers and develops in the field of analytical systems, web-systems, electronic document management systems and 1C. The company specializes in the design and development of information systems for insurance companies. The second key project is the development of an Internet engine for booking and purchasing air tickets.

Vitasoft needed a productive virtual server GPU for a project of. machine learning Prior to that, the company did not have very successful experience in renting equipment, so it was important to be able to test and ensure that the declared characteristics of the hardware were real. For testing, Cloud4Y offered servers Nvidia with Tesla V100 32GB that were able to completely close the needs of Vitasoft.

Virtual GPU servers with NVIDIA graphics cards are a good solution for working with big data, parallel computing, and training neural networks. The GPU server is capable of operating in 24/7 mode. The company gets the opportunity to immediately start training models thanks to pre-installed application packages PyTorch, TensorFlow, Keras, XGBoost, Scikit-learn, CUDA, OpenCV, Jupiter Notebooks and others. Cloud4Y also offers pre-installed libraries NumPy, Pandas and Scikit-learn to speed up the learning process. Access to graphics resources is possible from anywhere in the world and from any device.

After the test period, the provider's engineers deployed virtual machines on fast SSDs and with GPU servers. Additionally, the backup system was connected. This allowed the development company to get a stable and reliable virtual infrastructure, which can be quickly reconfigured depending on the requirements for the ML model.