CI/CD, infrastructure and operations from one senior engineer: designed, built and run by the person you talk to.
Scattered scripts, manual deployments and blind spots in monitoring become one platform: deployments without downtime, measured in daily operations.
10+
years of IT experience
8
service areas
1
direct point of contact
Services
From the first pipeline to calm operations: eight areas, each grounded in production practice. Book them individually or combine them.
Automated build, test and deployment pipelines, from commit to production. The same commit produces the same release, every time.
Deployments flow through Git: the desired state lives versioned in the repository. Argo CD or Flux continuously reconcile the cluster against it, with rollback by commit.
Metrics, logs and alerts that surface problems before customers notice. Dashboards that answer the 3 a.m. question: what broke, and where.
Stable Linux systems and automation of recurring tasks: cleanly scripted, idempotent and documented for handover.
$ ansible-playbook site.yml --check
ok=42 changed=0 failed=0 · idempotent
$ kubectl get nodes
3/3 Ready · all systems nominal
$ uptime
up 10+ years · still running.
Setup, tuning and hardening of relational and in-memory databases, including tested backup and restore strategies.
Infrastructure that lives in Git instead of in heads: versioned, repeatable and documented, for servers, networks and services.
Containerized workloads and Kubernetes clusters: from dockerizing existing applications to production-grade cluster operations.
Security as part of the pipeline: vulnerability scanning, clean secrets management and hardened systems instead of after-the-fact patches.
Architecture reviews, tooling decisions and workshops. The goal: your team runs and extends the platform on its own.
Anonymized excerpts from recent years. Detailed references and contacts available in a personal conversation.
Replaced manual release processes with multi-stage GitLab CI pipelines including automated tests, artifact management and rollbacks.
45 → 10 min
deploy time after the pipeline rebuild
0 ports
exposed to the outside; access only via VPN and YubiKey
Built a complete infrastructure with no publicly reachable services. The only way in is a VPN tunnel, the only key a YubiKey.
Built a highly available Kubernetes cluster on bare metal, including Helm-based deployments and clean environment separation.
0× downtime
for deployments in day-to-day operations
Introduced Prometheus, Grafana and Loki for an existing product team, including the alerting concept and runbooks.
h → min
time to root cause
< 30 min
to a tested recovery
Hardened a PostgreSQL cluster, established point-in-time recovery and automated regular restore testing.
Alongside client work I build products of my own, from first idea to architecture and day-to-day operations. A selection:
In production
Multi-tenant platform that brings provisioning, inventory (CMDB), monitoring and cost tracking for several companies together. Creates VMs on Proxmox and in the cloud, optionally fully encrypted with remote unlock via SSH.
In development
Renters document the vehicle condition in a guided photo walkaround in the browser, an ML model detects damage. The result is a cryptographically signed PDF report, stored in line with GDPR.
In production
Browser extension backed by a lean read-only API: shows prices, terpenes and details for medical cannabis flowers where patients already look. Data provenance stays strictly separated.
In development
A private budgeting PWA that pulls accounts from two banks via FinTS and PSD2. Field-level encryption, passkey-only login, reachable only inside a private network.
In development
Incremental game with its own engine, a prestige system and a bot that hunts down balancing issues automatically. More than 1,100 tests keep the mechanics stable.
In production
A browser extension reads the position, a Rust program analyzes it with engines like Stockfish and draws suggestions right on top of the board. Opening book and voice output included.
Open source
Go command-line tool that stitches tens of thousands of map tiles from an online game into a single image. Output as PNG, BigTIFF or DeepZoom pyramid, processed with libvips.
In development
Semi-automated production chain: LLM-drafted scripts, a synthetic voice, rendering with Remotion and FFmpeg, upload via API.
Behind bae.hr is one person with more than ten years of experience in IT: from classic system administration through automation to the architecture and operation of complete platforms.
No sales layer, no handovers, no project ping-pong: you work directly with the person who designs, builds and operates your infrastructure.
Contact
You will hear back within 24 hours on business days. Briefly describe what you need.
Available for new projects
Prefer to write directly?
[email protected]Your details are used exclusively to process your inquiry. See the privacy policy for details.