RAG for Legacy Systems: 7,432 Pages to 3s Answers
Production RAG for legacy systems: model-agnostic reranking validated across four LLM families. Real metrics, no vendor lock-in, 7,432 pages to 3s queries.
https://clouatre.ca/posts/rag-legacy-systems/
Production RAG for legacy systems: model-agnostic reranking validated across four LLM families. Real metrics, no vendor lock-in, 7,432 pages to 3s queries.
https://clouatre.ca/posts/rag-legacy-systems/
clouatre.ca
RAG for Legacy Systems: 7,432 Pages to 3s Answers | Hugues Clouâtre
Production RAG for legacy systems: model-agnostic reranking validated across four LLM families. Real metrics, no vendor lock-in, 7,432 pages to 3s queries.
Engineering VP Josh Clemm on how we use knowledge graphs, MCP, and DSPy in Dash
Dropbox’s VP of Engineering explains how the company built Dropbox Dash, an AI-driven cross-app search and knowledge system that uses indexing, knowledge graphs, and contextual reasoning to unify work content from many tools into a single, context-aware platform. The piece highlights engineering decisions about when to use indexed retrieval over federated approaches, how MCP tool calls a...
https://dropbox.tech/machine-learning/vp-josh-clemm-knowledge-graphs-mcp-and-dspy-dash
Dropbox’s VP of Engineering explains how the company built Dropbox Dash, an AI-driven cross-app search and knowledge system that uses indexing, knowledge graphs, and contextual reasoning to unify work content from many tools into a single, context-aware platform. The piece highlights engineering decisions about when to use indexed retrieval over federated approaches, how MCP tool calls a...
https://dropbox.tech/machine-learning/vp-josh-clemm-knowledge-graphs-mcp-and-dspy-dash
dropbox.tech
Engineering VP Josh Clemm on how we use knowledge graphs, MCP, and DSPy in Dash
Engineering VP Josh Clemm deep-dives into how we think about knowledge graphs, indexes, MCP, and prompt optimization using tools like DSPy.