mixashap

Dev Ops / Ml Ops Engineer

DevOps engineer with 8+ years in IT, 4+ focused on DevOps with building production infrastructure end-to-end — from bare Docker Compose to Kubernetes/Helm at scale. Took the 'Word Catcher' LLM product from zero to a registered, market-launched release; delivered enterprise CI/CD for large government and corporate clients. Specializes in containerized GPU/LLM workloads (vLLM, CUDA), full-stack observability, and Infrastructure as Code.

Experience: 4 years

Yearly salary: $40,000

Hourly rate: $30

Nationality: 🌏 Remote

Residency: 🇷🇸 Serbia

Experience

DevOps Engineer

TeDo (ex-PricewaterhouseCoopers Russia)

2021 - 2026

Accomplished market launch and patent registration of the 'Word Catcher' LLM product as measured by 5 days reduction in time-to-production, by designing a microservice architecture with vLLM in Docker and S3-based user data storage, deployable via both Docker Compose and Helm. Accomplished stable production serving of LLM workloads on GPU as measured by 99% uptime, by deploying and operating CUDA-based containers for vLLM and Ollama model serving. Accomplished 93% reduction in mean-time-to-detect incidents as measured by full-stack observability coverage across 5 services, by building monitoring and alerting with Prometheus, Grafana, Loki, and Zabbix, with automated reporting via Airflow, MS Teams, and email. Accomplished a successful production build of an undocumented legacy C++ project as measured by zero-regression compilation from source, by reverse-engineering the QlikView build dependencies and reconstructing the missing build pipeline.