The Future of Feature Stores and Platforms - MLOps Podcast 186
Offered By: MLOps.community via YouTube
Course Description
Overview
Dive into an in-depth discussion on the future of feature stores and platforms in this MLOps podcast episode featuring Mike Del Balso, CEO & Co-founder of Tecton, and Josh Wills, Angel Investor. Explore the evolution of machine learning infrastructure, from the creation of templates to intricate data handling techniques. Gain insights into Tecton's potential for sharing fraud and third-party features, and understand the critical role of quality data in AI systems. Delve into technical aspects such as optimizing models, addressing data pipeline challenges, and the importance of cohesive feature infrastructure in production environments. Learn about the birth of feature stores, their challenges, and their evolution, including the integration of Apache Iceberg. Discover the concept of dedicated query engines in feature platforms and the impact of LLMs on feature stores. Examine the role of vector databases, workflow templating efficiency, and even a gamification suggestion for Tecton. This comprehensive conversation covers a wide range of topics essential for understanding the current state and future direction of feature stores and platforms in the MLOps landscape.
Syllabus
[] Introduction to Mike
[] Takeaways
[] Features of the new paradigm of ML and LLMs
[] D. Sculley's papers
[] The birth of Feature Store
[] Data Pipeline Challenges Addressed
[] Operationalizing
[] Feature Store Challenges
[] Z access
[] Addressing Technical Debt Challenges
[] Real-Time vs. Batch Processing
[] Feature Store Evolution: Apache Iceberg
[] Feature Platform: Dedicated Query Engine
[] The bottleneck
[] LLMs, Feature Stores Overview
[] Vector databases
[] Workflow Templating Efficiency
[] Gamification suggestion for Tecton
[] Wrap up
Taught by
MLOps.community
Related Courses
Building Modern Data Streaming Apps with Open SourceLinux Foundation via YouTube How to Stabilize a GenAI-First Modern Data LakeHouse - Provisioning 20,000 Ephemeral Data Lakes per Year
CNCF [Cloud Native Computing Foundation] via YouTube Data Storage and Queries
DeepLearning.AI via Coursera Delivering Portability to Open Data Lakes with Delta Lake UniForm
Databricks via YouTube Fast Copy-On-Write in Apache Parquet for Data Lakehouse Upserts
Databricks via YouTube