From: Poland
Location: Pisz, PL
On Useme since 12 August 2026
About me
Process automation and data processing in Python | ESCP Business School | PL / EN / FR
I build scripts and integrations that remove repetitive, manual work.
WHAT I DO MOST OFTEN
– processing and cleaning data: CSV, Excel, deduplication, validation
– extracting and structuring data from websites and PDF documents
– API integrations and automations in n8n and Make
– generating reports and import-ready files
HOW I WORK
– a closed scope agreed before we start, with no drift along the way
– a written acceptance criterion — you know upfront what "done" means
– I state plainly what I do not promise, instead of letting you find out at delivery
– uncertain data is flagged, never guessed and never zeroed out
All my project code is public: github.com/Vic-glb
Stack: Python, pandas, openpyxl, API/REST, n8n, Make, SQL, TypeScript.
Categories
CV / Résumé
Jan 2025 - Now
Python Developer / Automation
Projekty własne
Building Python scripts and automations: data processing and cleaning (CSV/Excel), extraction from PDF documents, deduplication, validation, API integrations and n8n/Make automations. Public code: github.com/Vic-glb
Sep 2024 - Apr 2027
Student — Bachelor in Management
ESCP Business School
Studies at one of Europe's leading business schools. Programmes in Turin, Madrid and Berlin. Languages: Polish (native), French (native), English (fluent). Basics: Spanish and German (university courses), Chinese (3 years of study).
Portfolio
Problem: invoice data retyped by hand. Solution: extracts fields from PDFs, each carrying a status (found / check this / not found), plus NIP checksum and arithmetic controls. Result: errors are flagged, never silently corrected. 103 tests.
Web app (Next.js, TypeScript) for tracking contributions and settlements: who put in what, what each session returned, and what is still owed at any point in time. One test replays a full spreadsheet to the cent.
Problem: broken CSV/Excel files block every import. Solution: a cleaner that sorts each record into fixed, needs checking, or unusable — nothing is ever guessed. Result: you see exactly what is missing. 102 tests, public code.
Problem: monthly reports assembled by hand. Solution: an Excel generator with native charts, driven by a config file. Unreadable values are excluded with their reason, never zeroed or guessed. 81 tests.
Problem: nobody knows upfront how long fetching hundreds of thousands of records will take. Solution: a pilot mode that measures real throughput and projects the time for the full job, naming what limits it. Respects robots.txt, bypasses nothing.