Key components
Infrastructure should support both web systems and AI services.
- Proxmox cluster
- Kubernetes
- Docker ecosystem
- GPU inference servers
- resource monitoring
- CI/CD for deployment
Infrastructure
Cloud, DevOps and Infrastructure: Servers, Kubernetes, Docker, CI/CD and AI Inference
AI solutions need not only models, but also infrastructure: servers, containers, task queues, monitoring, inference services and repeatable deployment.
DevOpsKubernetesDockerProxmoxGPU inference
Context
Infrastructure as the base for AI projects
The laboratory has its own computing infrastructure and uses it for educational, research and applied tasks. This helps validate hypotheses and deploy prototypes faster.
A DevOps approach is important for any serious project: code should run consistently in test and production environments, and errors should be visible through monitoring.
Why it matters
Server infrastructure makes it possible to run models, APIs, databases, queues, analytics and education environments without depending on a single local computer.
Reliability
Serious systems need backups, logging, load control, orderly updates and an understanding of which services are critical.
Related work
Projects and research from the laboratory
Research Surgical Instrument Recognition System for St. Panteleimon Hospital The laboratory researched and developed a camera-based computer vision system for recognizing, training and digitally cataloguing thousands of surgical instruments.
Research Research and Development of a CRM System for Unbroken Hospital The laboratory researched and developed a demonstrational CRM system for a medical institution: patients, staff, wards, tasks, NK 025:2021 diagnoses, PDF reports, permissions and audit logs.
FAQ
Questions partners usually ask
Why do AI projects need GPU servers?
GPUs accelerate model training and inference for computer vision, NLP and other ML scenarios where CPUs may be too slow.
Is Kubernetes always required?
No. Kubernetes is useful for many services, scaling and stable deployment. For small prototypes, Docker Compose or a simpler setup may be enough.
Can resources be requested for educational tasks?
Yes, for educational tasks students can contact the laboratory with the goal, usage time and CPU or GPU resource needs.
Cooperation
Have a task in this direction?
The laboratory is ready to discuss research, prototypes and non-commercial projects with universities, laboratories, companies, hospitals and public institutions.