Johan Stemmet

Background

About Johan Stemmet

I’m a reliable Senior Linux & DevOps/Platform Engineer with 25+ years’ experience who builds automated, resilient infrastructure—and I’m now extending that into LLMOps and agentic AI. My core work spans multi-server Linux estates on VMware/Proxmox and HP servers (SLES/RHEL/Rocky/CentOS), CI/CD with Ansible/Terraform, and data platforms with ClickHouse/PostgreSQL/MongoDB. I built a custom bootable Linux OS + application ISO that cuts our customers' multi-server rollouts from ~4 weeks to ~3 days, and standardized PostgreSQL backups company-wide with pgBackRest (full/diff/incr, WAL, PITR, tested restores). I install and manage Apache NiFi → ClickHouse/PostgreSQL/MongoDB pipelines, keep services observable with Zabbix, and have achieved 2–5 years OS uptime on several production nodes. I’ve also recovered multiple service-critical, boot-failed servers (GRUB/LVM/kernel) remotely within hours, and wrote a company-wide hardware sizing calculator that right-sizes CPU/RAM/disk/LVM per server & app based on subscriber/load. Scale:~4–7 customers, each between 4–14 servers (≈16–50 total). I’m actively upskilling in LLMOps and agentic AI—building small RAG prototypes (PostgreSQL + pgvector) with LangChain and local inference (vLLM/Ollama), plus basic evaluation and guardrails—bringing the same platform rigor (IaC, CI/CD, observability, DR) to AI workloads. What I bring: delivery speed via automation, reliability through SRE discipline (runbooks, DR, pgBackRest), and a platform mindset that creates paved, self-service roads for teams. I enjoy automating noisy work, hardening systems, and mentoring engineers. Open to Linux support/Platform/DevOps/SRE roles and AI infrastructure/LLMOps opportunities.