M.Sc. Computational Cognitive Science @ University of Warsaw • NLP/ML Engineer • ex-Google SWE Intern • Research + Applied AI
I enjoy shipping useful tools, collaborating with researchers, and keeping things reproducible.

• Languages: Polish (native), English (C1), French (B1), German (A2), Russian (A2)
• Martial arts: Brown belt (1 Kyū) in Judo
• Hobbies: language learning, logic puzzles, travel

• Google — SWE Intern (GCP): Built and shipped a GCP Assistant in Python with the Agent Development Kit, registered on Vertex AI and deployed to Agentspace that orchestrates REST API calls, generates cURL commands, and supports human-in-the-loop steps to help engineers manage cloud resources and query docs.
• Institute of Computer Science, Polish Academy of Sciences (PAS):
  • PolEval — prepared large-scale document-layout datasets; manually corrected YOLO outputs and refined annotations (text, titles, headers, figures, tables, footnotes) for accessibility modeling.
  • Infostrateg — product matching/cleaning with robust regex rules; created gold-standard groupings for classifier training across markets.
  • ParlaMint — Python tools for NoSketchEngine + an interactive Colab (frequency plots, cross-corpus comparisons).
• Anex Poland: Python automations for Finance & Accounting — regex-based transaction classification, daily bank checks, error detection, reporting — and support for a Dynamics 365 migration.

• Languages: Python, Java, C++, R, SQL, OCaml
• ML / DL: TensorFlow, NLP, CNNs, transfer learning, data augmentation
• Data & Workflow: NumPy, Pandas, Jupyter/Colab, Weights & Biases, Excel
• Cloud: Google Cloud (Vertex AI, Agentspace)
• Other: Git, MATLAB, SPSS, RegEx

• Agent Simulation — Cognitive Processes Modelling II — Built a 2D agent simulation with a configurable evolutionary algorithm; modular engine, Pymunk physics, interactive Pygame loop, reproducible via config.txt.
  Tech: Python, Pygame, Pymunk • Repo

• GDP/Population/CO₂ Analysis Package — Python package to download, clean, and analyze macro datasets for selected countries.
  Tech: Python, NumPy, Pandas • Repo

• Polish Political Affiliation Classifier — Trained multiple TensorFlow models to predict party support from tweets; explored preprocessing variants and reported results.
  Tech: Python, TensorFlow, W&B, NLP • Report

• Symmetric Dependency Structure of Coordination: Crosslinguistic Arguments from Dependency Length Minimization
  Proceedings of TLT 2024, pp. 11–22, Hamburg, Germany. Association for Computational Linguistics.
  Read on ACL Anthology: https://aclanthology.org/2024.tlt-1.2/

• Natural Language Processing in TensorFlow — DeepLearning.AI — Credential
• Convolutional Neural Networks in TensorFlow — DeepLearning.AI — Credential
• Introduction to TensorFlow for AI, ML, and DL — DeepLearning.AI — Credential