Marek Cichy - Curriculum

My current goal

Career transition into a data scientist with focus on NLP from a Polish-Portuguese translation and interpretation role.

My background

Warsaw-based Portuguese language specialist with 8 years of experience. In this time I've:

• translated over 4000 pages, served as an interpreter for about 400 hours.
• created and managed Audioteka.es, the Spanish-speaking version of a Polish audiobook startup
• cofounded Mandioca, a Latin American graphic novel publisher.

Currently I continue self-learning of Python, Machine Learning in general and NLP in particular under Piotr Migdał's guidance. I've just finished the WEC Omdena (AI4Good) energy transition project in the roles of ML Engineer and Task Manager.

Portfolio:

External projects

• Omdena + World Energy Council project: AI for Energy: Transitioning toward a Sustainable Energy System in the role of ML Engineer
• Omdena + SexedPL's challenge: Analyzing the Importance of Sex Education With AI in the role of Junior ML Engineer

Private projects

• PT/BR Classifier, classifying Portuguese texts according to their variety (European or Brasilian Portuguese)
• NSFW-bot, Tweepy-based tweetbot consuming several image classification APIs
• Pyster, a script for automated posting on an invite-only forum.

Articles on ML and Python

• Collaborator Insights: Are Women Objectified in Polish Popular Music? NLP Applied.
• How I didn’t become a Data Scientist in 6 months
• Como criar um detector de sotaques de português com palestras TED
• NLP basics, hands-on: a language classifier deployed online in 3 steps
• How to set up an obscenely easy tweetbot with Python
• Does AI have a dirty mind, too?

Books and courses

• Data Scientist in Python learning path on DataQuest
• Python for Everybody by Charles Severance
• 19 DataCamp courses completed

Talks

- "I scraped 5000 TED talks and know my sambas from my fados" on ML in PL Conference 2019

Probable next steps

Upgrading the PT/BR Classifier:

• exploring the errors for patterns;
• adding basic XAI to the frontend;
• checking out an LSTM-based version of the classifier.
• broadening the corpus (crowdsourced movie subtitles, YT comments, etc.)

Writing about:

• PT/BR Classifier pt 2,
• My experience with collaborative challenges in Omdena

Reading:

• Introduction to Statistical Learning by Tibshirani et al. (under way)
• Deep Learning in Python by Francois Chollet