PyData Trójmiasto #39

We are very happy to invite you to 39th edition of PyData Trójmiasto meetup!

Where: Łużycka 8B, Gdynia, Tensor Z building.
When: 28th of January at 6PM

Registration: please register with your full first and last name. Don't forget to bring your ID when entering the event venue.
RSVP available until 6pm day before. In any urgent communication please let us know at kontakt@pydata-trojmiasto.pl

Parking: Free parking located along the street next to the building.

Agenda:

18:00 - 18:05 - Event boarding
18:05 - 18:10 - A few words about PyData
18:10 - 18:55 - „Lets sprinkle AI“ (responsibly). An approach of secure NLP implementation by Michael Rogocki
19:00 - 19:45 - Benchmarking Reasoning Models, Bielik-R case by Patryk Orwat
19:45 - Networking & snacks

About „Lets sprinkle AI“ (responsibly). An approach of secure NLP implementation by Michael Rogocki:

AI and Machine learning applications in high risk applications can be quite challenging. In this talk I would like to refer about NLP techniques in Python, the architecture, design and implementation of an anonymization framework and secure operationalization of NLP Models (LLM, Tok2Vec and Transformer) in a secure cloud environment. As an approach to "sprankle" AI solutions responsibly.

Michael Jan Rogocki is a Machine Learning Engineer and Data Scientist with over five years of experience developing data-driven software solutions. He is currently the Co-Founder at cm-opti, focusing on ML consulting and holds a Master of Sciences in Business Analytics, Düsseldorf Germany and B. Sc. in Logistics Engineering from TH Köln.
After working in predictive modeling for the logistics and retail sector, he has been recently involved in NLP and deep learning use-case implementations for the financial industry

About Benchmarking Reasoning Models, Bielik-R case by Patryk Orwat

As reasoning benchmarks evolve, the Bielik-R model development provides critical insights into the comprehensive evaluation of Polish language models. I will present how we developed a specialized platform for this task that features both a logic-based dataset and a rigorous analytical framework, potentially used to verify both RL with either ORM or PRM. The platform tracks key reasoning metrics, including token usage with output probability heatmap, circular reasoning detection, and efficiency scoring, specifically tailored for Polish NLP.
Patryk Orwat is a Principal Software Architect at Nokia with 15 years of software development experience across various sectors and organization sizes. As a technology enthusiast, recently contributed to the research and benchmarking efforts for the Bielik-R model.