The Stack Overflow Podcast Why build your own vector DB? To process 25,000 images per second
12 snips
Feb 7, 2025 Babak Bezad, Senior Engineering Manager at Verkada, dives into the cutting-edge realm of AI and image processing in video security. He shares insights on creating vector embeddings to handle massive data, utilizing technologies like YOLO and CLIP. The conversation explores the complexities of building a privacy-focused AI cloud and the importance of consumer control over features. Bezad highlights advancements in real-time anomaly detection and the integration of AI models to boost image processing accuracy, showcasing the future of security technology.
AI Snips
Chapters
Transcript
Episode notes
Diverse Customer Base
- Verkada's customer base is diverse, ranging from small businesses to large corporations.
- They serve gyms, retailers, schools, hospitals, and manufacturing facilities.
Image Processing at Scale
- Verkada processes 25,000 images per second, using AI models for object detection.
- These images are converted into vector embeddings and stored in a vector database.
Balancing Algorithms and Hardware
- Verkada combines clever algorithms and substantial hardware to achieve high processing speeds.
- Techniques like deduping and pre-processing optimize GPU utilization.