Fractional Analytics Engineer

Job listing We’re looking for an experienced fractional analytics engineer to help keep our BI reporting reliable during a team transition. This is hands-on work across Looker, LookML, dbt, SQL, data-source migrations, validation, and troubleshooting. You should be comfortable tracing a metric from its source through the data model and dashboard, finding why numbers differ, and documenting what people can safely rely on. Current needs include migrating reports away from legacy data paths, validating historical continuity, maintaining executive and operational reporting, investigating stale or inconsistent data, and reviewing technical changes before they reach production. The work will vary week to week. We need someone who can take a well-scoped reporting task, work independently, raise risks early, and explain findings clearly to technical and business partners. Responsibilities Build and update dbt models, LookML, Explores, Looks, and dashboards. Debug joins, grain, filters, attribution logic, and metric definitions. Support data-source and warehouse migrations, including AWS replication and legacy connector transitions. Validate migrations using row counts, time-window comparisons, known records, freshness checks, and dashboard reconciliation. Maintain reporting continuity during marketing-data cutovers, including historical data, backfills, and monitoring. Investigate stale sources, missing data, dashboard regressions, and differences between reports. Maintain executive pacing and other recurring business reports. Review dbt and Looker pull requests, tests, and CI results. Document assumptions, open risks, and operating procedures. Give short, clear updates that business partners can understand. Must-have skills Strong SQL and dimensional data-modeling experience. Production experience with dbt. Production experience with Looker and LookML. Experience debugging metric differences across source, model, and dashboard layers. Experience validating data-source or dashboard migrations. Familiarity with Git, pull requests, CI, testing, and code review. Experience with Redshift, Snowflake, BigQuery, or a similar cloud warehouse. Strong written communication and sound judgment when evidence is incomplete. Nice-to-have skills AWS Aurora-to-Redshift Zero-ETL or similar replication experience. Fivetran migration or connector-management experience. Iterable, Segment, or other marketing-event data experience. Source-freshness monitoring, historical backfills, and incident response. Ecommerce, marketing attribution, revenue, or pacing-report experience. Python for small reporting automations. Experience writing concise runbooks and data-quality notes. Engagement terms Hourly contract for up to 30 hours per week. Hours may vary with the backlog and reporting incidents. Recommended initial term of eight weeks, with an option to extend. Flexible schedule with an agreed amount of overlap with the internal team. Weekly priority check-in and short written progress updates. Work goes through tracked tasks, pull requests, and review before production changes. No general on-call expectation. Any time-sensitive coverage is agreed in advance. Start with one paid, scoped task to confirm fit before expanding access or scope.