Data Engineering
Pointline Data Lake — Architecture & Design
Pointline data lake architecture and design guidance. Use when: (1) designing new tables, modules, or subsystems for the pointline data lake, (2) evaluating design trade-offs (encoding, partitioning, storage, schema evolution), (3) writing or reviewing ExecPlans or research proposals, (4) assessing change risk levels (L0/L1/L2) for PRs, (5) reviewing code for PIT correctness, determinism, or invariant violations, (6) planning new vendor integrations or data source onboarding, (7) making architectural decisions about schema, storage, or pipeline extensions, (8) understanding why existing design choices were made (function-first, fixed-point, quarantine-over-drop, no backward compatibility, SCD2), (9) extending the module dependency graph without introducing cycles.