On demand
Claude Code for Data Engineering
Data engineers are well acquainted with fire drills. A pipeline breaks, but no one knows why or what's impacted downstream. A user reports data staleness, but it takes hours to track down the upstream failure. A KPI value changes, and the reason is buried under 3 simultaneous PRs.
This session covers how Anthropic's data engineering team hands much of that reactive work to Claude — freeing engineers for the higher-level architecture that prevents these issues in the first place. We'll walk through the failure modes to watch for and demo the Claude Code workflows we use every day.
What you’ll learn
How to design a Claude-powered triage workflow that aggregates operational signals, filtering out noise and elucidating action items
How to analyze a pipeline failure with Claude as the investigator and orchestrate class-wide remediation through Claude
How to create automated review bots with Claude that encode known pain points, so violations are caught at PR time
How to construct requirements and design documentation through iterative dialogue with Claude, structured so both human teammates and Claude sessions can act on it