Uploaded August 2026 | Updated September 2026, 2 weeks ago
with Sr. Product Marketing Manager Maddie Presland (Backblaze) and VP of Product Alison Hill (Pixeltable)
You have a carefully curated collection of videos that you want to use as a dataset for model training and fine-tuning. But the raw media files require complex and expensive processing to create data useful for training, and you cannot know all the steps you'll need to run before you get started. How do you actually transform raw media files to training-ready data? Instead of standing up 5+ services and stitching them together, you can use Pixeltable to turn large video files into usable, searchable chunks and extracted frames — and keep all your inputs and outputs in Backblaze B2. Join Sr. Product Marketing Manager Maddie Presland (Backblaze) and VP of Product Alison Hill (Pixeltable) for a practical overview of setting up a production-ready media processing workflow that works at scale.
Join Backblaze experts and AI integration partners as we explain the importance of each phase, how high-performance storage factors into keeping AI datasets both secure and accessible, and how to build a new data pipeline from scratch or get more information from the data you already have using open source tools that integrate with Backblaze B2.
with Sr. Product Marketing Manager Maddie Presland (Backblaze) and VP of Product Alison Hill (Pixeltable)
You have a carefully curated collection of videos that you want to use as a dataset for model training and fine-tuning. But the raw media files require complex and expensive processing to create data useful for training, and you cannot know all the steps you'll need to run before you get started. How do you actually transform raw media files to training-ready data? Instead of standing up 5+ services and stitching them together, you can use Pixeltable to turn large video files into usable, searchable chunks and extracted frames — and keep all your inputs and outputs in Backblaze B2. Join Sr. Product Marketing Manager Maddie Presland (Backblaze) and VP of Product Alison Hill (Pixeltable) for a practical overview of setting up a production-ready media processing workflow that works at scale.
Join Backblaze experts and AI integration partners as we explain the importance of each phase, how high-performance storage factors into keeping AI datasets both secure and accessible, and how to build a new data pipeline from scratch or get more information from the data you already have using open source tools that integrate with Backblaze B2.
