How to Deploy AI Solutions on Custom Servers and Accelerate System Response Time by 80%
Client
CyberNet AI — a developer of AI solutions for banks, telecommunications companies, and government agencies.
Objective
Migrate to reliable and powerful servers within Uzbekistan, minimize downtime, accelerate product development, and improve IT system response times.
Solution
Results
Client Profile
CyberNet AI is an international developer of speech AI solutions for the financial sector, telecommunications, and government agencies. The company’s geographic presence covers Central Asia, North America, and the MENA region, with offices in 6 countries: Uzbekistan, Kazakhstan, Kyrgyzstan, Egypt, USA, and Mexico.
Having entered the Uzbek market in 2021, the company is confidently growing and expanding its client base. To date, its portfolio includes over 50 implemented AI projects across various industries. The primary customers for CyberNet solutions are banks, marketplaces, telecom operators, and retailers.
One of the key advantages of the company’s voice AI bots is the high quality of speech technologies, enabling smooth dialogue with users in the local language without long pauses or a foreign accent.
Why CyberNet AI Chose Servercore
This level of service did not meet the company’s needs, so the CyberNet AI team began searching for a new technology partner in the Uzbek market. A key task for the AI developer was to ensure flexible scaling of cloud capacity to accelerate the release of new products and improve customer service.
The key factors in choosing Servercore were:
«To develop our AI solutions in Uzbekistan, we were looking for a local cloud provider with reliable infrastructure and SLA guarantees. Our goal was to accelerate development and ensure a high level of service for customers. When we learned that Servercore allows deploying custom configurations for specific tasks and flexibly scaling resources, we decided to work with them».
Key Client’s Objectives
Solution
Outcomes and Plans for the Future
CyberNet AI successfully migrated to Servercore’s custom cloud servers within 4 weeks. The company deployed the backend of its key products — voice AI agents — on the leased resources. The company utilizes a high-load microservices architecture with ML models in production and a large volume of UDP traffic.
Cloud servers with over 99.98 % uptime guaranteed by SLA ensure stable operation of the AI services, which has increased their availability and improved IT system response speed by 80 %. The use of custom configurations with flavors allows the company to flexibly manage CPU load and reduce costs.
Due to the high scalability of Servercore’s cloud solutions, the AI service developer was able to instantly increase leased resources for new client projects and significantly increase the speed of product development. Furthermore, the company is satisfied with the convenient management achievable in just a few clicks with the Servercore control panel.
As of the time of writing, CyberNet leases over 70 virtual machines at Servercore with more than 200 CPUs. The company’s project utilizes approximately 500 GB of RAM for efficient and fast work with ML models.
CyberNet AI’s plans include expanding its product portfolio, notably by deploying an LLM model, launching speech analytics, and chat-bots. For these new ambitious AI projects, the company will require servers with GPUs, and it is prepared to choose Servercore solutions in Tashkent.
