The specific signals an experienced analytics engineer hiring panel grades on during the eight-second scan.

• Summary names the mart and the dbt scope

  'Owns the growth + finance marts in dbt (220 models)' beats 'analytics engineer.' Mart + model count is what panels scan for.

• Semantic-layer tool named

  dbt Semantic Layer, Cube, Looker LookML, Mode datasets. Generic 'semantic models' parses as junior.

• Downstream exposure surfaced

  Number of downstream consumers (dashboards, reports, models). Exposure-driven SLAs are the AE-specific signal.

• BI tool partnership named

  Looker / Mode / Hex / Metabase / Sigma — the BI layer the analytics engineer hands off to.

• Data-quality tests on critical metrics

  dbt tests on metric primitives, freshness SLAs, ownership coverage. AEs are graded on metric reliability.

• Business-metric ownership (where applicable)

  MRR/ARR/cohort/retention metric definitions owned. The senior AE-specific signal.

1. 1

   Open with the mart and the dbt scope

   A senior AE summary names the marts and the model scope: 'Analytics engineer at a Series C SaaS; owns the growth + finance marts in dbt (220 models, 38 downstream dashboards).' Mid: 'AE on the data team; owns the growth mart (62 models, 14 dashboards).' Entry: 'Recent grad; shipped a 38-model dbt project on a real e-commerce dataset, deployed to a public Hex dashboard with weekly active viewers.'

2. 2

   Quantify with model count + exposure

   dbt model count per mart, downstream dashboard / report / consumer count, freshness SLA compliance, metric drift, self-service query deflection. These are AE-specific units.

3. 3

   Name the semantic layer + BI tool

   dbt Semantic Layer + Looker LookML is one common stack. Cube + Hex is another. Mode + custom-SQL is a third. Naming your primary semantic-layer + BI tool signals AE fluency.

4. 4

   Surface metric-definition ownership

   AEs at the senior level own metric primitives. Pattern that works: • Name the metrics (MRR, ARR, churn, cohort, activation). • Name the semantic-layer tool. • Quantify the reconciliation work (alignment with GL, with PM, with finance). • Quantify the drift outcome over a sustained window.

5. 5

   Close with dbt OSS / community / publication

   High-signal closing items: merged PRs to dbt-core, dbt packages (dbt-utils, dbt-expectations), or recognized analytics open-source. Coalesce talks. Locally Optimistic guest posts. Substantial published metric-definition or modeling work.

Each follows the [verb] [object] [number] structure hiring managers grade against. Copy them as a starting point, swap in your own numbers, and read the annotation to understand why each one works.

• dbt scope

  Own 220 dbt models across 4 marts (growth, finance, product, ops); model-execution time fell 38% via incremental materialization + clustering keys; freshness SLAs documented per model with ownership.

  Why it works: Model count, mart count, execution-time delta, and the SLA + ownership pattern.

• Metric ownership

  Authored the MRR + ARR + cohort-retention metric definitions in dbt Semantic Layer; ran a 4-week reconciliation pass with the finance team to align the warehouse metrics with the GL; metric drift on finance dashboards fell to zero across 6 quarters.

  Why it works: Names the metrics, the tool, the reconciliation work, and the longitudinal outcome.

• BI handoff

  Built the Looker LookML translation layer for the growth mart (38 Looks + 12 dashboards); analyst self-service queries rose from 240/month to 1,400/month, reducing AE-team ad-hoc query load by 78%.

  Why it works: Names the BI tool, artifact count, and the self-service deflection outcome.

• Data quality

  Built 38 dbt tests across critical gold-tables (freshness, uniqueness, referential integrity, business-rule); freshness violations dropped 84% in 6 weeks; ownership documented per test with on-call rotation.

  Why it works: Test count, categories, freshness outcome, ownership pattern.

• Framework

  Authored the growth-mart cohort framework (cohort-by-acquisition-source, cohort-by-plan, cohort-by-feature-adoption); now powers the weekly growth-review meeting and the quarterly retention deep-dives.

  Why it works: Names the cohort dimensions, the cadence it powers, and the meeting/review surface. AE framework work that institutionalizes signal.

• Migration

  Migrated the growth mart from spreadsheet-driven KPI tracking to a dbt + Looker pipeline; replaced 38 manual weekly reports; cycle time on KPI changes dropped from 5 days to 4 hours.

  Why it works: Names the before/after surface, report count, and a cycle-time outcome.

• Tooling

  Authored the team's dbt style guide (model naming, ref pattern, materialization defaults, test coverage minimums); adopted across 3 data sub-teams; PR review time on model changes fell from 2 days to under 6 hours.

  Why it works: Style-guide adoption, cross-team scope, PR review time outcome.

• Lineage

  Built the data-lineage layer (dbt-docs + Marquez) exposing upstream/downstream for every model; surfaced 14 orphaned models (no downstream consumers) — reclaimed via deprecation runbook.

  Why it works: Names the lineage stack, the orphan-detection count, and the deprecation runbook outcome.

• Strategy

  Co-authored the company's metric tree (4 layers: input → driver → KPI → north-star); now serves as the canonical map for cross-team metric questions and onboarded into the data-team interview rubric.

  Why it works: Names the metric-tree structure, the canonical-map role, and the interview-rubric adoption — multi-team scope.

• Open Source

  Two merged PRs to dbt-labs/dbt-utils — one extended date_spine to support week-aligned partitions; one closed a unicode-collation bug in deduplicate.

  Why it works: Named package, PR count, two technical descriptions.

• Mentorship

  Mentored 2 data analysts transitioning to AE focus; both shipped sole-owner mart additions (one growth, one ops) within 6 months.

  Why it works: Mentorship transition (analyst → AE), timeframe, mart-level deliverable.

• Entry-level

  Built a 38-model dbt project on a real e-commerce dataset for a university capstone; deployed to a public Hex dashboard; 4,200 weekly active viewers from the campus community.

  Why it works: For entry-level AE candidates, this kind of E2E shipped dbt project with public users is high-leverage.

Same intent, two phrasings. Read why the right column lands on the keep-pile and the wrong column doesn't.

Patterns our writers see most often when reviewing analytics engineer resumes — each one disqualifies candidates faster than weak experience does.

• Mistake

  Generic 'analytics engineer' opener without naming the mart.

  Fix

  Name the mart you own (growth, finance, product, ops). Mart is the AE scope signal.

• Mistake

  dbt model claims without count or exposure.

  Fix

  Surface model count + downstream dashboard/consumer count. AE scope is graded on exposure, not just SQL volume.

• Mistake

  Metric work without naming the metric or the semantic-layer tool.

  Fix

  Name MRR / ARR / churn / cohort + the dbt Semantic Layer / Cube / LookML.

• Mistake

  BI partnership without naming the tool.

  Fix

  Name Looker / Mode / Hex / Metabase / Sigma by exact product. BI fluency is AE-specific.

• Mistake

  Listing both data-engineering and analytics-engineering as equal expertise.

  Fix

  AE and DE are distinct in 2026. Tilt your resume toward the role you're applying for.

• Mistake

  No dbt mentioned.

  Fix

  dbt is universally the AE substrate in 2026. If you don't use dbt, name what you use (SQLMesh, Dataform) precisely.

• Mistake

  Two-page resume below 7 years experience.

  Fix

  One page. AE hiring moves fast.

• Mistake

  Listing every BI tool you've touched.

  Fix

  Name your primary; tier the others. Listing 6 BI tools reads as junior.

Strong verbs lead strong bullets. Replace generic openers (worked on, helped with, was responsible for) with the specific verb that matches what you actually did.

ATS pipelines weight your Skills section as a structured list. Include 15-25 of the items below if they match your experience — not soft skills.