talk-data.com
People (22 results)
See all 22 →Activities & events
| Title & Speakers | Event |
|---|---|
|
Panel Talk: Meet the Professionals
2025-07-01 · 15:00
David Lapsanschi
– Head of Data Science
@ Evozon
Panel featuring four professionals from industry and academia sharing experiences, insights, and guidance for students and beginners in data science. Moderated by the Head of Data Science at Evozon. Topics include career paths, programming languages to focus on, building your first portfolio, and landing your first data science role. |
|
|
Managing MLflow for LLM projects
2025-07-01 · 15:00
Learn how to set up MLflow for LLM tracing and monitoring in this practical session. We’ll walk through the basics of capturing execution traces from language model applications and explore how MLflow can help you track and debug your LLM workflows. Using straightforward LangChain examples, you’ll see how to implement basic tracing functionality to gain better visibility into your model’s behavior and performance. An introduction for data scientists and ML engineers who want to add observability to their language model projects. We will also go into managing the lifecycle of experiments, runs and traces. |
|
|
PyData Cluj-Napoca: Meetup #24
2025-02-13 · 16:00
🎉 Kick Off the Year with PyData! First Meetup of 2025! 🐍🚀 New year, new insights, same awesome community! 🎊 Join us for our first PyData meetup of 2025 this February—an evening of great conversations, knowledge-sharing, and networking with fellow data enthusiasts. Whether you're a seasoned pro or just starting your data journey, this is the perfect place to connect, learn, and have fun! Don't forget to bring good vibes and warmth for a social and informative evening! (+🍕🍔🌯🌮) ---------------------------------------------------------------------------------------------------- "Beyond the ordinary — unconventional approaches to time series classification" by Andrei Lucian Established techniques such as ANOVA, ARIMA, DBScan, and XGBoost have demonstrated effectiveness in certain contexts. However, I propose exploring unconventional approaches that leverage novel mathematical concepts. This exploration is not necessarily about achieving success, but rather about gaining valuable insights through experimentation. ---------------------------------------------------------------------------------------------------- "Building better chatbots with Knowledge Graphs" by David Lapsanschi Chatbots powered by Retrieval-Augmented Generation (RAG) are great at pulling information from vast datasets, but they often stumble when it comes to understanding the bigger picture. They miss the subtle connections between ideas, fail to clarify ambiguous queries, and sometimes dish out responses that feel... a bit flat.Enter knowledge graphs: a tool for making sense of relationships between entities. Think of them as a web of interconnected concepts that let chatbots "see" how things are related, not just what they are. By integrating RAG with knowledge graphs, we can give chatbots the power to reason, connect dots, and deliver answers that make sense even in complex conversations.In this talk, we’ll explore how these concepts work together to overcome the limits of standalone RAG. We'll explore together when and why knowledge graphs can be a better solution that can unlock smarter, more dynamic chatbots that go beyond surface-level responses.With some real-world examples and practical insights, we'll show you what we learned in the last year of actively building RAG and knowledge graph solutions. More importantly, we'll go over the mistakes we made and how they helped us better understand the posibilities and limitations of the current technology. ---------------------------------------------------------------------------------------------------- NumFOCUS Code of Conduct https://numfocus.org/code-of-conduct |
PyData Cluj-Napoca: Meetup #24
|