Scalable GenAI in Production
summary
Concrete claims and concrete build details: a working Netflix generative search beta and a deployed Gemini Enterprise demo that connects Drive SharePoint and Jira ground the talk; practical stack pieces include vector DBs RAG pipelines and an orchestration/agents layer; measurable engineering decisions cover latency versus cost tradeoffs capacity planning and LLM judges for scalable evaluation; predictions claim agentic interfaces as default and small specialized models plus multiagent orchestration as the competitive moat; five production imperatives and a data flywheel framework are given.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 2:55Scalable GenAI stack
- 5:44Grounding and RAG for LLMs
- 7:52Scaling GenAI lessons
- 14:53GenAI enterprise maturity
- 19:09GenAI predictions
- 22:18Production GenAI imperatives
- 24:03Q&A
related talks
23mScaling GenAI Applications and Agentic AI for Business Competitive Differentiation
22mLearning to Apply Generative AI at Enterprise Level
26mbuilding reliable AI systems in production
18mNavigating the growth frontiers of generative AI
20mThe secret path to practical generative AI in the enterprise
38mPanel Discussion: AI Needs a Foundation That Never Fails
15mChallenges in moving Gen AI apps to production
17mGenerative AI: Keeping it real