Accelerate experiment design with Amazon Bedrock
summary
LaunchDarkly shows that integrating Amazon Bedrock (Titan) and a vector database into the experiment builder generates test variations from historical experiments, flag metadata and hypothesis text, and suggests the correct primary metric. The team moves from prototype to production in weeks and ships a working feature in months, automatically creating feature flags, emoji-tuned headlines and semantic search to explore the parameter space more efficiently and reduce reliance on data scientists.
AI-written summary · about
chapters (11)
- 0:00Introduction to speaker
- 2:25Experimentation basics
- 6:10The role of parameterization
- 10:08Data-driven product strategy
- 15:09Building experiment workflows
- 21:56Leveraging generative AI
- 27:11Integrating LLMs in product
- 30:08Generative AI demo
- 34:08Benefits of AI integration
- 37:17Using Amazon Bedrock
- 41:46Future strategy and closing
related talks
18mAccelerate development with the Engineering Development Hub
43mBeyond Chatbots: How Amazon Bedrock Agents Drive Enterprise AI
39mModel choice in Amazon Bedrock
40mBuilding an AWS solutions architect agent with Amazon Bedrock
39mEvaluate and choose models in Amazon Bedrock (AIM204)
01h 00mSecurely accelerating generative AI innovation
31mBuilding a Data Experimentation Platform Without a Platform Team
30mAccelerate AI to prod