About
Senior software and data engineer with 5+ years building production data pipelines, backend services, and autonomous AI-agent systems in Python. Strong across the modern data stack (dbt, Airflow, BigQuery), software engineering practices (Terraform/IaC, CI/CD, TDD), and LLM application development.
Write Python in core application code, not just pipelines. Proven record of automating business processes, cutting infrastructure cost, and translating finance and business requirements into tested, version-controlled systems. Champion AI-assisted development — sharing reusable conventions and tooling that lift team velocity.
Experience
Jan 2026 — Present
Data Engineer · Booksy
- Build and maintain production data pipelines on GCP (BigQuery + Airflow) as part of the data engineering team, supporting a marketplace with 260M+ annual bookings across 5 markets.
- Designed dbt models for SaaS revenue recognition aligned with GAAP / ASC 606, partnering with Finance to turn accounting requirements into tested, version-controlled SQL; outputs feed audit-ready monthly reporting.
- Drove dbt test coverage from ~22% to ~97% across the production model catalogue — 10 to 144 models with tests, over 1,500 assertions added.
- Managed GCP infrastructure as code with Terraform across 3 environments (dev, prod, sandbox); added CI checks to catch issues before release.
- Author and maintain shared coding conventions and Claude Code skills for the team; run workshops that improve code quality and delivery velocity.
Oct 2023 — Dec 2025
Software Engineer (Python / AI) · Morele.net
- Built autonomous AI agents (Claude / LLM-based) automating business and customer-service workflows; automated 70%+ of complaint resolution across ~5,000 weekly cases, cutting operational costs ~20%.
- Designed reusable agent "skills" that execute end-to-end tasks — including one driving a live Selenium browser for autonomous website integrations and scraping.
- Built an internal platform that lets non-technical staff configure, deploy, and monitor 300+ production bots; used daily by 10 senior staff.
Jun 2022 — Oct 2023
Data Analyst → Data Engineer · Occubee
- Built and maintained PySpark pipelines processing POS and receipt data for ~500k SKUs per client, forecasted and ordered daily or weekly with compute and cost optimised to business needs.
- Onboarded retail clients onto Occubee's AI-driven replenishment and allocation platform; one client released ~PLN 4M in capital previously frozen in excess inventory while growing revenue.
- Contributed to fashion-industry forecasting for leading retail brands — model-colour-size demand allocation across store networks.
- Proposed and shipped new features that improved forecasting and replenishment accuracy.
May 2020 — May 2023
Python Developer · Freelance
- Designed and shipped end-to-end ML applications (e.g. stock-price prediction): Selenium data collection, feature engineering, model training, and deployment behind Django/REST APIs with JavaScript front-ends.
- Built invoicing automation that auto-generates hundreds of correction documents daily, replacing a fully manual workflow and saving ~4 hours of operator work daily.
May 2021 — Jun 2022
Analyst → Process Automation · PwC (Financial Crime Unit)
- Began as an AML/KYC analyst running financial-crime checks and data validation procedures.
- Transitioned to the process improvement team: delivered Python automations that replaced manual data ingestion workflows for analysts — moving from spreadsheet-driven data collection to automated pipelines and boosting team throughput.
Projects
Wedding guest media uploader. Guests scan a QR code and upload photos, video, or voice messages straight to a Google Drive folder. Offline-resilient upload queue via IndexedDB; shared gallery with infinite scroll and lightbox; admin panel for moderation. Full Playwright e2e test suite.
Python system ingesting GPW market data via APIs and Selenium scrapers. Signal generation pipeline runs as a LangGraph StateGraph — fetch, compute indicators, evaluate rules, LLM analysis, dispatch — with configurable provider (Ollama, OpenAI, Anthropic) and broker (paper or XTB). Being extended to automate trade execution end-to-end.
Serverless GCP pipeline that fetches PM2.5, PM10, O₃, and NO₂ measurements from the OpenAQ API via a Cloud Function, validates them, and writes CSV outputs to Cloud Storage on a schedule.
Scrapes EBIT and amortisation figures for WIG20 companies from biznesradar.pl and computes EBITDA — data normally gated behind a premium subscription. Handles compute-cost-optimised fetching, bank exclusions, and numeric coercion.
Skills
Education & More
Education
MSc Management of Investments and Real Estate
University of Gdańsk · 2022
Erasmus exchanges in Finland and Spain.
Certificates
- PCPP1 — Certified Professional Python Programmer (OpenEDG)
- Python TDD: Test-Driven Development and CI/CD
- Machine Learning A–Z: Python & R in Data Science
- 100 Days of Code: Python Pro Bootcamp
- SQL – MySQL for Data Analytics and Business Intelligence
Languages
- EnglishC1 / C2
- SpanishB2
- PolishNative
Interests
- Machine learning & applied AI
- Algorithmic trading & stock markets
- Football