Explore the challenges and trends in Generative AI production in this insightful podcast episode featuring Verena Weber, a Generative AI Consultant. Delve into key topics such as model size, context window breakthroughs, the shift from unimodality to multimodality, and the impact of EU AI Act regulations. Learn how Gemini 1.0 and 1.5 exemplify these trends. Gain valuable insights from Weber's diverse experience in data science, NLP, and deep learning, including her time at Amazon's Alexa team. Discover practical applications of GenAI in business, strategies for improving AI systems like Alexa, and the importance of data quality in AI development. The discussion also touches on promoting diversity in tech and creating inclusive environments for underrepresented individuals in the field.
GenAI in Production - Challenges and Trends - Lecture 224
Overview
Syllabus
[] AI Quality Conference
[] Verena's preferred coffee
[] Takeaways
[] Ski Person of Influence
[] Verena's background in the last 5-10 years
[] Tech Evolution: Rapid Transformation
[] Working at Amazon and key challenges
[] Research-inspired suggestions
[] AI Updates Impact Workflows
[] Alexa Query Distribution Analysis
[] Innovative Solutions for Alexa
[] Robust T5 Data Prompting
[] Audio Data Quality Challenges
[28:21-] Zilliz ad
[] Alexa data transcription and data cleaning
[] Considering needs, costs, and complexity
[] ChatGPt is not ideal for classification
[] Comparison of model building using TF, IDF
[] Struggle to boost diversity in conference speakers
[] Creating safe environments helps underrepresented individuals participate
[] Wrap up
Taught by
MLOps.community
