Leipzig, Saxony, Germany
Most companies leaving VMware trade one dependency for another — a hyperscaler, a licensing model, an AI API that ships their data abroad. Synohm exists for the ones who want out for good. I design and deploy sovereign infrastructure stacks: everything runs on open-source software and hardware you own, in your rack or your datacenter. Not a slide-deck reference architecture — a working platform I've built, hardened, and documented end to end: • Virtualization & storage — Proxmox VE + Ceph, highly available, provisioned as code (OpenTofu + Ansible), rebuildable from a single config file • Identity & trust — self-hosted SSO with enforced MFA (Authentik), your own private certificate authority (step-ca), secrets that never touch a repo • Workloads — CIS-hardened RKE2 Kubernetes with GitOps (Argo CD), backed by your own git remote — no GitHub dependency • On-prem AI — tiered local LLM inference (OpenAI-compatible, so your tools just work) with the governance layer enterprises actually need: deterministic checks gate, models only advise; autonomy is propose-only and explicitly mode-gated; every decision is auditable • Operations — nightly verified backups, central auth-log observability, break-glass paths for every SSO dependency, documentation generated from machine-readable sources of truth so it can't silently drift That last part is the point. Running Ollama on a server is easy. Running local AI inside an auditable, GDPR-clean, EU-AI-Act-ready platform — with the same operational rigor you'd demand from any production system — is what Synohm delivers. Deployments are fixed-scope and fixed-price. You get a defined outcome, full documentation, and a handover your own team can operate.
Synohm delivers sovereign infrastructure and on-prem AI platforms as fixed-scope, fixed-price deployments — built entirely on open-source software and customer-owned hardware. The product: a fully documented reference platform, developed in-house covering virtualization (Proxmox VE + Ceph HA cluster), identity and PKI (Authentik SSO + private CA), hardened Kubernetes (RKE2, Cilium, GitOps via Argo CD + self-hosted Forgejo), and a tiered local-LLM inference layer with enterprise governance: propose-only autonomy, deterministic gates with LLM-advisory-only checks, full audit trails. The method: everything as code. Infrastructure declared in OpenTofu, configured by Ansible, documented from machine-readable YAML sources of truth with automated drift detection — the entire platform rebuilds from one configuration file. Typical engagements: • VMware exit: assessment → Proxmox/Ceph target architecture → migration → handover • Sovereign AI enablement: on-prem LLM inference integrated into existing infrastructure, GDPR- and EU-AI-Act-aligned • Platform build-out: identity, PKI, Kubernetes, backup/DR on open source
Supporting many cultural events, organizing and assisting in Events - Animal Rights Congress Leipzig / GSO Leipzig - Planning, building and setting up a Server Lab Environment based on redundant layered Network Stack with Mikrotik and Proxmox. Research about and use of LLM Models, locally and distributed.
Rollout of Key Projects, VMware SME, RCA - Mission Critical Production Systems, Impact Mitigation of unresolved issues SSH, Ansible, Confluence, Python, Project Management
daily ops, change requests, incident handling, project involvements, VMware SME CMS, RDP, SSH, GIT, Ansible, ITSM
• ESXi/vSphere storage and network support log/dump analysis, storage/network analysis, HCL checks, WCP
• troubleshooting vSphere 5.5 - 7.0 log analysis, vSphere AD and SSO integration, vSphere Certificates, Postgres/SQL, shell scripting
• monitoring and maintenance of systems and business performance • leading daily performance conference calls with international teams • managing access and interaction routing profiles Avaya Telephony, Genesys CME, Tableau, CRM, IAM, shell scripting, incident management
• second level support for all Apple products, consumers and small businesses log analysis, network/storage troubleshooting, user account management
• first level support for all Apple consumer products