From Quick Wins to Revolutionising Productivity & CX with GenAI
summary
Claims a working pipeline that vectorizes entire videos with specific timestamps into DataStax Astra DB, using ViT‑B/32 for frames (about 1,536 sampled frames) and Whisper for audio chunks, stores embeddings and metadata, and exposes a Langlow no‑code flow plus REST curl endpoint. Retrieval uses cosine similarity and hybrid search; a demo pulls the winner moment at 2:17 but sometimes misses exact endpoints because of sentence chunking and start‑time capture. Open‑source model runs locally; solution is built and demonstrably runnable.
AI-written summary · about
chapters (9)
- 0:00GenAI and semantic search
- 3:58Data Stacks and Cassandra
- 8:57Vector search basics
- 14:41Video demo with search
- 17:26Audio processing challenges
- 22:24Building Langflow AI
- 25:35Live demo and results
- 27:43Technical stack and deployment
- 28:27Q&A
related talks
17mAI/ML demo derby
33mEnter the Brave New World of GenAI with Vector Search
27mUsing GenAI with Audio Data for Driving Better Customer Experience
57mEnter The Brave New World of GenAI with Vector Search
46mEvent Streaming and its Potential for the GenAI World
18m: Workforce Efficiency with AI-Driven Vector Search
58mAI-Enhanced Big Data: Integrating Private LLMs and Vector Databases
19mBuild Real time, AI powered Apps