DataTalks x Sparks Meetup: From Magic to Mastery with LLMs


Details
📢 DataTalks x Sparks Innovation Summit: From Magic to Mastery with LLMs!
Our next DataTalks meetup will dive into the fascinating world of LLM evaluation and creativity, featuring deep-dive talks, good company, and a relaxed atmosphere—hosted at the Polly Bar in Tel Aviv, as part of the Sparks Innovation Summit! 🎉
📍 Location: Polly Bar, Rothschild 60, Tel Aviv-Yafo
🚘 Parking: Beit Hadar (Ahuzot HaHof)
🛡️ Bomb Shelter: 20 seconds from venue
***
🗓️ Agenda
🍕 18:30–18:45
State of What Art? Measuring the Sensitivity of LLMs to Prompt Variations
Speaker: Moran Mizrahi
Tiny changes, big impact—this lightning talk explores how LLMs can behave unpredictably with small prompt tweaks, and how we might evaluate their stability more effectively.
🔶 18:50–19:35
Cooking Up Creativity: Enhancing LLM Creativity through Structured Fusions
Speaker: Moran Mizrahi
What happens when you blend generative AI with culinary data? Discover how structured fusion techniques unlock novel applications for creative tasks—like designing new recipes with ChatGPT!
🦭 19:35–20:00
Mingle Break
Grab a drink, chat with fellow ML folks, and enjoy the vibe.
🔷 20:00–20:50
From Magic to Science: The GenAI Way
Speaker: Dr. Nataly Kuritz
GenAI may feel like magic—but it’s time to engineer trust. This interactive session explores how to evaluate generative AI systems practically and transparently, integrating responsible AI practices into your workflow.
***
👩🏫 About the Speakers
Moran Mizrahi recently completed a PhD at the Hebrew University, focusing on computational creativity in NLP. Her research explores how we can teach machines to be meaningfully inventive.
Dr. Nataly Kuritz is an AI strategy advisor and data science expert, specializing in building trust and transparency into GenAI systems through thoughtful evaluation and process design.
***
We hope to see you there for great talks, thoughtful discussion, and a drink or two! 🍻

DataTalks x Sparks Meetup: From Magic to Mastery with LLMs