Skip to main content
Prompt recording

Lyft on MCP and data context for any AI surface

With AI, analysis can happen anywhere. Devon Mittow, who leads data platform work at Lyft, keeps metrics and context in a tool-agnostic layer so analysis can start in Claude, pass through an MCP, and land in Hex. He and Hex co-founder and CTO Caitlin Colgrove dig into the next wave of data engineering: less about moving data between tools, more about structuring context for whatever surface the question comes from.

Lyft on MCP and data context for any AI surface

What you'll learn

  • Why Lyft keeps metrics and context in a tool-agnostic layer
  • How analysis flows from Claude, through MCP, and into Hex
  • What the next wave of data engineering looks like when questions can come from any surface