Systematically Improving Your RAG - jxnl.co
jxnl.co
Systematically Improving Your RAG - jxnl.co
- Requirements (or constraints) : What does success look like? What can we not do?
- Methodology : How will we use data and code to achieve success?
- Implementation : What infrastructure is needed in production?
Real-time Machine Learning For Recommendations
Here is the plan they developed: The people write down every customer-driven change they make. They count the aggregate number and then categorize them to spot trends. When they see trends, they review the internal systems and structure to see if they are aligned with where the market is going. They discuss among themselves what it will take to do
... See moreJames A. Belasco • Flight of the Buffalo: Soaring to Excellence, Learning to Let Employees Lead
MLEs are happy to delegate experiment tracking and execution work to ML experiment execution frameworks, such as Weights & Biases 3 , but prefer to choose subsequent experiments themselves. To be able to make informed choices of subsequent experiments to run, MLEs must maintain awareness of what they have tried and what they haven’t (Lg2 calls ... See more
Shreya Shankar • "We Have No Idea How Models will Behave in Production until Production": How Engineers Operationalize Machine Learning.
Google search and Wikipedia in particular are great examples of how the internet has enabled immense scale and speed: every crawlable page on the web, every sufficiently noteworthy topic, etc, returned to you within a second. But choosing the extreme end of scale is at odds with the key property of personal knowledge management: relevance. A scient... See more