Explore a 56-minute MLOps Community Mini Summit featuring Pavol Bielik, CTO of LatticeFlow, and David Garnitz, CTO of VectorFlow, discussing model blind spot discovery and handling unstructured data at scale. Delve into the challenges of building high-quality AI models, including data biases and systematic failures. Learn how LatticeFlow empowers ML teams to deliver resilient models through data and model diagnosis. Discover VectorFlow's approach to creating reliable vector embedding pipelines for connecting raw data to LLMs. Gain insights on the AI Bill, open source vs. closed source models, and controlled open source strategies. Connect with the speakers and MLOps community through provided links and explore related resources for further learning.
Model Blind Spot Discovery for Better Models and Ingesting Chaos - Mini Summit #2
Overview
Syllabus
[] Introduction to Pavol Bielik & David Garnitz
[] Pavol Bielik - Navigating Model Blind Spot Discovery for Better Models
[] David Garnitz - Ingesting Chaos
[] AI Bill
[] Open source vs closed source models
[] Open source success
[] Controlled Open source
[] Critical takeaway
[] Wrap up
Taught by
MLOps.community
Related Courses
-
Getting Started with Mistral
-
Fresh Data and Smart Retrieval: Milvus and Jina CLIP Explained - MLOps Mini Summit 7
-
Building a Production Scale, Totally Private, OSS RAG Pipeline with DBRX, Spark, and LanceDB
-
Supercharge Your LLM Applications with RAG
-
Want High Performing LLMs? It's All About Your Data - Importance of Quality Data in AI Development
-
Alignment is Real - Fine-Tuning vs. Prompting in Domain-Specific Tasks
