- Repetitive work?
- I take it personally.
- Broken workflows?
- I rebuild them.
- Six tools, no system?
- I make them talk.
- Weeks of waiting?
- Gone.
- The result?
01 · where it started
Origin
I study digital marketing with a heavy lean toward data and code: a B.A. at SRH Fernhochschule, expected in 2027. Somewhere between the campaign decks and the tracking sheets I noticed the thing that would not let me go: how much of every project was waiting.
Waiting for an agency. Waiting for a handoff. Waiting for a spreadsheet to be updated by hand.
02 · the first script
The turn
At some point I stopped building the tenth slide and built the first script instead. That was the turn: friction is not a fact of life. It is a to do.
a script writing itself
- shortcutsparse what nobody wanted to parse
- pythonscrape, sort, file
- jiraremove the manual loop for good
- done. friction is a to do.
03 · what I built
Proof
Every project below was built the same way: in stages, each one closed by a written review, with results measured against a target before anything was called done.
case 01
Community Signal Pipeline
- Concept and build
- Python, SetFit classification, Docker, n8n
- 2026
- Problem
- Customer objections were being discussed publicly at scale, but only ever read anecdotally. Decisions leaned on gut feeling because nobody could count what people actually said.
- Outcome
- What used to be read anecdotally is now counted, in a single overnight run. Delivered as a working prototype, not a slide deck.
- Build
- An automated pipeline that collects public discussions, removes personal data before anything is stored, and classifies every post by sentiment and objection category. Each figure links back to its anonymised source, and the classifier was tuned against a reference set labelled by hand until its accuracy cleared 70 %.
How it was run: in stages, each closed by a written review, with a completeness check of what was collected against what was expected before any figure left the pipeline.
case 02
A launch-ready funnel for a new online course
- Concept, build, decision paper
- HTML/JS, serverless API, EU hosting
- 2026
- Problem
- The course needed proof of demand before the product existed: leads, a pricing model, a legal footing, and data that stays in the EU.
- Outcome
- From concept to launch readiness, with a decision paper the partner can act on.
- Build
- Landing page, a lead backend with double opt-in, campaign attribution, pricing and partner model, legal texts, and a move of the whole stack to EU providers.
How it was run: one status board for every workstream, a decision paper per open question, and a design check that measures interface states instead of eyeballing them.
case 03
Abstracta: research reading, automated
- Design and build
- Python, PyMuPDF, pdfplumber, web front end
- 2026
- Problem
- Reading scientific papers for AI-assisted research meant copying text out of PDFs by hand, with columns, footers and tables scrambled on the way.
- Outcome
- A folder of papers becomes clean Markdown in one command. Its landing page, which tells its story as you scroll, was the interface study behind this site.
- Build
- An offline converter that reconstructs reading order, strips headers and footers, keeps tables intact, names files by year, author and title, and runs on a whole folder at once. No API, no cloud, 100 automated tests.
How it was run: 100 automated tests before the first release, and a landing page built as a study of exactly the kind of scroll story you are reading now.
three of many, chosen for what they changed
04 · today
Now
Today I am a working student in social media marketing at BMW Motorrad, with my degree running alongside. And I build my own tools: Python pipelines, AI classification, the web. Building the tool is the easy half. The harder half is a team that actually wants to use it.
Change that sticks is the whole job.
see the loop, build it, win the team, it sticks
- OBSERVEsee the loop
- AUTOMATEbuild it
- ADOPTwin the team
- CHANGEit sticks
Open to full-time roles, working-student and internship positions, and a bachelor thesis in marketing automation, AI and digital processes. Open to conversations.
05 · where I'm from
Roots
I grew up on one train line. At home in Freising from the age of five, every school I went to was there, and along the way I trained as a certified 3D designer by distance learning, because I needed to know how things are built. Since 2020 I have also been self-employed with Simplify Music. Two semesters of biotechnology taught me I'd rather work with people than with pipettes. Then the line runs into Munich: an HR marketing internship at Theo Müller, a year at LOBECO's community desk for automotive brands, and since February 2026 a working student in social media marketing at BMW Motorrad. I still cook the way I build: exact, tested, then shared.
The eight stops
01 · Unterschleißheim · 2003–2008
The first five years
Gone by five. Everything after that is Freising.
02 · Freising · since 2008
Home, still. And every school
Primary, then Realschule with French, then FOS in the social branch: two internships with kids, one in an inclusive after-school club. Since 2020 also self-employed as Simplify Music: social media strategy, podcasts and video for clients.
03 · everywhere (distance learning) · 2021–2022
Learning how things are built
Certified 3D designer, by distance learning at ILS. Then the Fachabitur, 2023.
04 · Freising-Weihenstephan · 2023–2024
Two semesters of biotech
Enough to learn I would rather work with people than with pipettes.
05 · everywhere (distance learning) · since 2024
Digital marketing, from wherever the laptop is
B.A. Digital Marketing at SRH Fernhochschule, expected 2027. The campus is the S1.
06 · Freising · Oct 2024–Jan 2025
HR marketing intern at Theo Müller
Marketing for people, not products: employer branding content and the numbers behind it.
07 · Munich (LOBECO) · 2025–2026
Into the city
Community management at LOBECO for international automotive brands.
08 · Munich (BMW Motorrad) · since Feb 2026
Social media at BMW Motorrad
Working student in social media marketing.
06 · where this goes
In a few years I want to be the person who leads AI process automation and digital change across teams, not because of the tools, but because of how people adopt them.
If that sounds like a conversation worth having: let's have it.
