| Home | https://cusy.io |
| GitHub | https://github.com/cusyio |
| https://www.linkedin.com/company/cusy-gmbh | |
| Meetup | https://www.meetup.com/Cusy-Meetup/ |
Today, we took a closer look at the new OpenAI models and training methods. They seem promising, but their local use still needs to prove itself: https://cusy.io/en/blog/openai-open-weight-models.html
‘Single occurrences that cannot be reproduced are of no significance to science’ wrote Karl Popper in 1935 in The Logic of Research. This has not changed to this day. What is new is that research data and research software must be managed sensibly. To do this, you must not silently rely on certain resources and development environments. Changes to your data and software can be tracked and team collaboration can be facilitated.
In today’s digital landscape, maps have become essential components of countless applications and services, from navigation and logistics to social platforms and data visualisation. But for too long, the field has been dominated by a handful of companies whose services, while powerful, come with significant drawbacks: Usage quotas, tracking requirements, styling limitations, and recurring costs that can quickly skyrocket as applications grow.
Data Scientists nutzen zunehmend LLMs (GitHub Copilot, LLama, Claude etc.) um schnell einen Prototypen für ihre Datenanalyse zu bauen. Ihre Erwartungen, was sie innerhalb kurzer Zeit mit einem LLM erreichen wollten, sind fast grenzenlos. Häufig ist das Programmieren mit LLMs jedoch schwierig und wenig intuitiv. In diesem Vortrag möchte ich etwas von unserer über zweijährigen Erfahrung mit LLMs zur Datenanalyse mit Python an euch weiterzugeben.
cusy GmbH
Veit Schiele 🔜 @FrOSCon
