Atlassian & CMDB depth
Jira, Jira Service Management, Assets — administration, workflow automation, and CMDB architecture covering users, applications, servers and contracts. AQL, REST APIs, OAuth, webhooks: the whole ecosystem, hands-on.
PRAGUE · HYBRID / REMOTE · OPEN TO EMPLOYMENT & CONSULTING
Jira Service Management & CMDB specialist. Integration engineer. AI-augmented builder. The person you call when the problem doesn't fit any single department.
David Porteš
I joined Správa železnic — the Czech national railway infrastructure manager — straight out of school as a junior C# developer. Five years later I run the CMDB in Jira Service Management, having grown through internal development and full Atlassian administration along the way.
But titles are not the whole story. Over time I became the person people came to when a problem didn't belong to any particular department: I translate technical problems into business language and back, and I know how systems work — and how the people around them work, too.
My scope is deliberately wide, not narrow. I won't out-code a senior C# specialist, but I know a lot of tools, I make them talk to each other, and I keep the team running while I do it — access, onboarding, coordination, the unglamorous glue that projects actually run on.
Jira, Jira Service Management, Assets — administration, workflow automation, and CMDB architecture covering users, applications, servers and contracts. AQL, REST APIs, OAuth, webhooks: the whole ecosystem, hands-on.
Python and FastAPI services, SAP and SharePoint data migrations, CI/CD pipelines, Docker, nginx, systemd. If two systems need to exchange data reliably, I build the bridge — with tests.
Coordinated the day-to-day of a 12-person dev team — access, onboarding, helpdesk, paperwork. Interface point between an 11-person ITSM department and the rest of a large organisation. I make the boring parts run so the experts can be experts.
Správa železnic, state organisation — Czech national railway infrastructure manager · Prague
Apr 2026 — present
Sep 2024 — Apr 2026
Sep 2021 — Sep 2024
Work systems are described without internal names. Home-lab systems run publicly on my own infrastructure — including this site.
Work · national rail infrastructure manager
Routine Jira group management and CMDB changes flowed through admins as tickets — slow for everyone.
A FastAPI web portal with Azure AD single sign-on and server-to-server OAuth into Jira: role-based self-service for group management, company & contract records, and webhook-driven writeback into the CMDB with a full audit log. Deployed to RHEL 9 with GitLab CI/CD, nginx and TLS.
Live in production · dev + prod environments · gitflow pipeline
Work · national rail infrastructure manager
Certificate records in the CMDB drifted from reality; renewals were tracked by hand.
A two-phase sync pipeline: a morning run parses certificates and posts a human-approval preview; an evening run writes approved changes to Jira Assets. Detects SAN drift, auto-derives validity status, and closes the loop on the service-desk ticket.
346 automated tests · approval-gated writes
Work · national rail infrastructure manager
Asset data lived in SAP; the CMDB was fed by hand.
An import pipeline from SAP exports into Jira Assets — the first integration between the two systems there, replacing the manual process entirely.
Manual process eliminated
Work · national rail infrastructure manager
No usable backup story for the cloud Jira, Confluence and Assets data.
A backup tool for the entire Atlassian cloud environment, designed and implemented from scratch.
Designed & built solo
Work · national rail infrastructure manager
A legacy internal system in SharePoint had to move to a new application without losing its data.
Full data migration: SQL export, cleaning and transformation, reload into the target system.
Legacy system retired
Home lab · justporo.cz
A family of four (plus picky kids) deciding what to watch every evening.
A family media assistant: swipe-based discovery, per-person taste profiles, a kid-safe advisor with rating filters, and recommendations generated by a local Ollama LLM. Integrates Radarr, Sonarr, Jellyfin, TMDB and ČSFD. CI/CD with per-PR preview deployments.
353 automated tests · PR preview deploys · in daily family use
Home lab · justporo.cz
Mornings spent checking feeds, two mailboxes, a calendar and Jira separately.
A daily digest service that pulls RSS, Gmail, Outlook calendar and Jira, then has a local LLM write one readable morning briefing.
One briefing instead of five inboxes
Home lab · self-hosted
Family apps need real infrastructure: TLS, deployments, backups, CI — not a laptop under the desk.
A self-hosted platform running ~25 Docker services behind nginx with Let's Encrypt: media stack, self-hosted git with CI runners, local LLM inference, Home Assistant integration, and a nightly AI ops agent that reports server health to my phone.
~25 services · public subdomains · nightly AI ops report
Home lab · hiring.justporo.cz
A CV should prove the claims it makes.
The page you're reading: FastAPI + Jinja2, bilingual content from YAML, a PDF CV rendered server-side from the same data — built with Claude Code in an agentic workflow and deployed through the same pipeline as everything above.
Data-driven CV · EN/CZ · self-hosted
An honest self-assessment — depth where it counts, breadth everywhere else.
I build with agentic AI tooling daily — Claude Code, MCP integrations, and local LLMs on my own hardware. Certified in autonomous agentic development (ICT Pro, 2026). The point isn't the buzzword: it's that I ship more, with more tests, than I could alone — and I know exactly where the tools stop and engineering judgment starts.
So neither of us wastes a call — this is the shape of role I'll say yes to:
A role that combines hands-on engineering with coordinating people and processes — I already do the informal version; I want the formal one.
ITSM, CMDB and integrations at bank / telco / corporate scale — thousands of CIs, real data flows, real stakes.
Based in Prague; hybrid preferred, remote works.
Full-time role, or consulting engagements alongside — both are on the table.
Available as contract work alongside my day job. If your Jira hurts or your CMDB is a spreadsheet, I can help.
Greenfield setup, workflow redesign, permission & scheme audits, automation.
Object model design, AQL, import pipelines, keeping the CMDB honest.
Python bridges between Jira and anything with an API — SAP, Azure AD, monitoring, in-house systems.
Practical LLM automation for IT operations — local or cloud, with engineering judgment included.
Rates on request — write me with your problem, not a job description.
I'm based in Prague, open to hybrid or remote work, and available for both employment and contract consulting. I'm employed and not desperate — but for the right problem, I'm very interested. Let's talk.