jarvis
JavaScriptJ.A.R.V.I.S.: my personal AI assistant with more than 120 tools — from calendar and research to deployments.
Details →View on GitHubSpecialist Support Engineer · Berlin
Daniele
Michelin
$
For 18 years I have been keeping cloud platforms running — today I connect classic infrastructure with AI automation. I make systems reliable, and the people around them happy.
Open to new opportunitiesCloud infrastructure · technical support / customer engineering · AI automation · Berlin
18+
years cloud & support
~15.000
Nextcloud instances
~150.000
HiDrive users
~200
tickets per week
$ whoami
daniele.michelin — Specialist Support Engineer · Cloud Storage & Collaboration
$ systemctl status daniele-michelin
● daniele-michelin.service — active (running) · uptime 18+ years
cloud-storage [ OK ] ~15,000 Nextcloud instances at 2nd/3rd level
collaboration [ OK ] ~150,000 HiDrive users
identity [ OK ] SAML · OIDC · M365
automation [ OK ] Python · Bash · Docker · n8n
ai-lab [ OK ] LLM · RAG · voice agents
$ cat next_task.txt
A new challenge where cloud and AI grow together.
Interactive terminal — type help or press ⌘K
01 · Numbers
Numbers that speak for themselves
18+
Years in cloud & support
~15.000
Nextcloud instances supported
~150.000
HiDrive users looked after
~200
Tickets per week
120+
Tools in my AI assistant
02 · Profile
From support ticket to AI automation.
I started at STRATO — today IONOS — in Customer Care back in 2008. Today I work as a Specialist Support Engineer at 2nd and 3rd level: where tickets end up once they stop being simple. My focus: Cloud Storage and Collaboration — Nextcloud, HiDrive, S3 Object Storage.
Along the way I built a second passion: AI that actually takes work off people's plates. I build RAG knowledge systems, run local language models and develop automations that give a full ticket board room to breathe again.
My belief: good technology is not the kind people talk about — it's the kind that simply runs. Preferably followed by a dry "it just works" from the other side.
- Location
- Berlin, Germany
- Role
- Specialist Support Engineer · STRATO / IONOS
- Languages
- Deutsch · Italiano · English
- Focus
- Cloud Storage · Collaboration · AI automation
- Status
- open to new opportunities
03 · Skills
What I keep running.
No buzzword bingo — these are the services I own every single day:
nextcloud.service
running
Enterprise-grade operation & troubleshooting — from sync conflicts to migrations, including direct database and system fixes on faulty instances. → SSO errors in Nextcloud
hidrive.service
running
Sync & share for ~150,000 users: diagnosing failures, securing performance, recovering data.
storage.service
running
S3 Object Storage & storage systems: lifecycle, backups, data integrity, latency and connection issues.
identity.service
running
SAML, OIDC, MFA, M365: integrations, SSO failure patterns, access concepts. → 8 SAML/OIDC errors in support
support.service
running
Third-level support: owning complex incidents and customer escalations, prioritising a high ticket volume in Jira — with AI-assisted triage.
automation.service
running
Python, Bash, Docker, n8n: log analysis by script, from manual labour to stable pipelines. → J.A.R.V.I.S.
ai.service
running
LocalAI, Whisper, RAG, agents: self-hosted AI that takes on real tasks. → LLM-Wiki (RAG)
Tech stack → Skill graph
- Cloud & storage
- NextcloudHiDriveS3 object storageSync & shareDesktop sync clients
- Identity & access
- SAMLOIDCSingle sign-onMulti-factor authenticationMicrosoft 365
- Linux & operations
- LinuxsystemdBashDockerPleskGit
- Data & migration
- Database analysisMigrations & updatesData integrityLatency analysis
- DNS, mail & TLS
- DNSTLS certificatesSPF · DKIM · DMARCHTTP / reverse proxy
- Automation
- PythonFastAPIn8nREST APIs
- AI
- LLMsRAGAI agentsLocalAIWhisper
- Support
- 2nd/3rd levelRoot cause analysisEscalation managementLog analysisJiraTicket prioritisationAI-assisted triageDocumentation
04 · Projects
Built because they didn't exist yet.
Selected projects from my workshop — all open source on GitHub.
LLM-Wiki
Python · JSRAG knowledge system: ask questions of your own wikis and documents — answers with sources included.
Details →View on GitHubjob-hunter
SwiftSelf-hosted job application automation for macOS: find jobs, score them, track them — no spreadsheet chaos.
Details →View on GitHubzimage-turbo-pipeline
PythonAI pipeline for image and video assets: generated locally, post-processed automatically.
Details →View on GitHubacestudio-mac
SwiftNative macOS interface for local AI music production with ACE-Step.
Details →View on GitHubfamiglia-os
HTML · JSThe digital home: calendars, tasks and family organisation in one place.
Details →View on GitHub05 · Lab
Work in progress — behind closed doors.
Not everything that runs shows up on GitHub. My own server is a playground, too. What's taking shape there right now:
Sofia
private repoMy personal voice agent: speech in, action out. LiveKit, local Whisper and TTS models, my own tools — fully self-hosted.
Knowledge Graph
private repoNeo4j + RAG: I turn my own knowledge into a searchable graph — the answer to "where did I read that again?"
Autopilot
private repon8n, monitoring, smart triggers: automations that keep my back free — from system alerts to the family calendar.
Media Studio
private repoA pipeline for AI images and videos: prompt in, finished asset out — on my own hardware, no external cloud.
Agent Office
work in progressMy office, staffed by agents: tasks in, results out — orchestration, memory and quality control on my own hardware.
OPSCOPE
work in progressRead-only Bash toolkit for diagnosing Linux servers: ten modules from network to mail, symptoms get correlated, and a finished report comes out at the end — built from everyday support work.
Watch the demo →Local AI receptionist
private codeAI receptionist that answers the phone or replies in the browser, books appointments and speaks Italian and German — listening, voice and brain run locally on a Mac. Tested with simulated calls, not yet with a real Twilio number.
Details →The rest stays private — but that's exactly what I love talking about. Ask me.
06 · Career
18 years, one direction: reliable.
2023 – today
AI & Automation
LLMs, RAG, agents: I bring AI into daily work — mine and that of the systems I support.
2018 – today
Specialist Support Engineer — 2nd/3rd level
Cloud Storage & Collaboration: Nextcloud, HiDrive, S3 Object Storage. When things get complicated, I step in.
2016 – 2018
Telekom Marketplace — B2B support (powered by STRATO)
Technical support for Telekom's Business Marketplace: B2B customers and partners, operated by STRATO.
2008 – 2016
STRATO (now part of IONOS Group) — Customer Care
Entering the cloud world: hosting, domains and the first steps with sync & share.
2001 – 2008
DHL — Logistics & Operations
Structure, pace, customer focus: the foundation for everything that followed.
Tools & reference
Free for everyday admin work.
Checks in your browser, error messages explained, commands to copy — no sign-up, no tracking.
Paste a log line: meaning, causes, commands — AI-assisted.
Open →PHP, Nextcloud, Debian & co: what still gets updates.
Open →Latest CVEs for Nextcloud, PHP, nginx, Synology, Plesk.
Open →Counts the DNS lookups of your SPF record against the limit of 10.
Open →MX, reverse DNS, SPF, DMARC and MTA-STS at a glance.
Open →Expiry, chain and names of a TLS certificate.
Open →136 error messages: meaning, checks, fix.
Open →occ, dig, journalctl, openssl, nginx, systemctl, Docker and Git — searchable, with a copy button.
Open →64 terms and 31 ports explained briefly.
Open →07 · FAQ
Quick answers.
$ Who is Daniele Michelin?
Daniele Michelin is a Specialist Support Engineer at STRATO (IONOS Group) in Berlin. He has worked in cloud storage and collaboration at 2nd/3rd-level support for more than 18 years and builds his own AI and automation tools.
$ What does he specialise in?
Cloud storage & collaboration — managed Nextcloud (around 15,000 instances), HiDrive (around 150,000 users) and S3 object storage — plus identity (SAML, OIDC), root-cause analysis and escalation management. Tech stack: Python, Bash, Docker, n8n, LLMs and RAG.
$ Is he open to new opportunities?
Yes — roles where cloud infrastructure, technical support / customer engineering and AI automation meet, based in Berlin.
$ Which languages does he speak?
German, Italian and English.
$ What has he built?
Among others J.A.R.V.I.S. (a personal AI assistant with 120+ tools), LLM-Wiki (a RAG knowledge system with citations), job-hunter (job-application automation for macOS) and acestudio-mac (local AI music production). All open source on GitHub.
$ How can I contact him?
By e-mail at info@daniele-michelin.com, via LinkedIn or the contact form on this page.
08 · Contact
Let's talk.
Open to new challenges in the cloud and AI space — or just a good conversation. I usually reply faster than a 1st-level ticket.
Status: open to new opportunities · Berlin
In IT today