Models, sources, seeds, tests, snapshots, macros, exposures, docs, CI/CD, lineage.
Before starting, we run a 1-minute tech check — microphone, ambient noise, connection. If your setup isn't good enough, the test is fully refunded.
Prove your dbt skills in 15 minutes: models, tests, macros, incremental strategies, and CI/CD — not just a buzzword on your resume.
The Plume dbt badge puts your real-world dbt knowledge to the test through a structured AI-led oral exam. The conversation covers the full dbt lifecycle: staging/intermediate/marts model architecture, materialization choices (view, table, incremental, ephemeral), generic and singular tests, SCD2 management with snapshots, Jinja macro authoring, packages like dbt_utils and dbt_expectations, schema.yml documentation and exposures, and CI/CD integration with slim CI and state:modified artifacts. The exam is calibrated for practitioners who use dbt in production, not just people who completed a tutorial.
What makes this badge more credible than a self-declared LinkedIn skill is the conversation itself. The AI examiner asks you to walk through real projects: a refactor of a legacy model layer, a test that caught a data anomaly in production, or an incremental strategy choice (merge vs insert_overwrite) driven by cost and volume constraints. It also probes the boundaries of the tool — when dbt is not the right answer, and why you might reach for SQLMesh, Dataform, or a Spark pipeline instead. After the exam, Claude Opus reads the full transcript and generates a score from 0 to 100 with a certified proficiency level.
This badge is built for data engineers, analytics engineers, and senior data analysts who work with dbt regularly — on dbt Core or dbt Cloud — and want objective proof of their level to land a new role, close a freelance contract, or make an internal case for a promotion. It's especially relevant if you work with BigQuery, Snowflake, Databricks, or Redshift and want to show you go well beyond a basic SELECT.
Here are the concrete dimensions the AI examines during the 15-minute oral.
Structuring a dbt project across staging, intermediate, and marts layers, applying consistent naming conventions, and managing dependencies through ref() and source() to maintain a clean, auditable lineage.
Choosing between view, table, incremental, and ephemeral based on cost, freshness, and volume constraints, and configuring merge or insert_overwrite strategies with unique_key and partition_by settings.
Writing generic tests (not_null, unique, accepted_values, relationships) and singular tests, leveraging dbt_expectations for advanced assertions, and wiring test results into CI to catch regressions before they reach production.
Building custom macros to factor out repetitive SQL logic, using Jinja if/for blocks, leveraging dbt_utils functions like generate_surrogate_key and union_relations, and understanding the compilation cycle.
Configuring snapshots to track slowly changing dimensions using timestamp or check strategies, and understanding the automatically generated dbt_scd_id, dbt_updated_at, and dbt_valid_to columns.
Integrating dbt into GitHub Actions, GitLab CI, or dbt Cloud Jobs, using manifest.json artifacts for state:modified runs, and orchestrating jobs with Airflow, Dagster, or Prefect.
Writing model and column descriptions in schema.yml, generating the dbt docs site, and defining exposures to trace how transformed data flows into BI tools like Looker, Metabase, or Tableau.
Identifying when dbt is the wrong tool — complex Python transformations, streaming use cases, extreme volumes — and making a reasoned case for alternatives like SQLMesh, Dataform, or a native Spark pipeline.
Final scoring is performed by Claude (Anthropic), which reads back the full transcript and applies this weighted criteria grid.
Precision on core dbt concepts: ref(), source(), materializations, incremental strategies, snapshots, Jinja macros, and packages. Candidates should use accurate terminology and demonstrate understanding of dbt's internal mechanics.
Quality and specificity of the projects you describe: DAG size, target warehouse, constraints faced, decisions made. Strong candidates cite real numbers, deliberate tradeoffs, and lessons learned from actual failures.
Knowledge of naming conventions, staging/intermediate/marts separation, lineage and source management, and project organization that scales to team collaboration and multi-environment deployments.
Ability to describe a working dbt CI/CD pipeline: automated tests, slim CI with state:modified, environment management (dev/staging/prod), and orchestration tooling choices.
Ability to identify dbt's limits, compare it with alternatives like SQLMesh or Dataform, and speak to recent developments such as native unit tests (v1.8), model contracts, model versions, or the dbt Core/Cloud/Fusion split.
A Plume session takes about 20 minutes, from tech check to badge delivery.
Your mic and connection are tested automatically. Make sure you're in a quiet spot. The entire exam is voice-based — no screen sharing, no live coding, no IDE required.
The AI invites you to give a quick intro and describe your most recent or most complex dbt project: warehouse type, DAG size, your role on the team. This calibrates how deep the next questions will go.
The AI examiner digs into 4 to 6 themes: model architecture decisions, materialization choices, tests that caught real issues, macro and Jinja usage, CI/CD setup, and your take on dbt's limits and recent evolution. Every answer can trigger a follow-up question.
You get a chance to add anything you didn't cover — a specific use case, an opinion on dbt Fusion or native unit tests. The AI closes the session and confirms the analysis is underway.
Claude Opus analyzes the full transcript and generates your score (0-100), your level (Novice/Proficient/Advanced/Expert), a detailed report per criterion, and the public URL of your badge ready to share.
Your score out of 100 translates into a level a recruiter can grasp at a glance.
You understand dbt's core concepts — models, ref(), dbt run, dbt test — but haven't shipped a project to production yet. You can follow tutorials and grasp the declarative SQL transformation logic, but advanced materializations, macros, and CI/CD are still unfamiliar territory.
You use dbt in production on at least one project. You can configure sources, write generic tests, manage seeds, and set up simple snapshots. You're starting to write macros and layer your models, but complex incremental strategies and CI/CD pipelines are not yet second nature.
You own the dbt architecture for a team project. You choose materializations deliberately, write custom Jinja macros, leverage dbt_utils and dbt_expectations, and have set up a CI/CD pipeline with slim CI. You document your models and exposures, and you can debug production run failures confidently.
You design large-scale dbt architectures on warehouses like Snowflake, BigQuery, or Databricks. You're fluent with model contracts, model versions, custom packages, and cross-project ref. You follow dbt Core's development closely and know exactly when to reach for SQLMesh, Dataform, or a Python-native pipeline instead.
No degree or years of experience required to take the badge. Here are the profiles it makes the most sense for.
You use dbt every day and want objective proof of your level to negotiate a senior title, a salary bump, or a lateral move — without sitting through a 3-hour technical panel at every company.
You come from a classic ETL/pipeline background and are adopting dbt for transformations. The badge validates that you've internalized dbt's conventions and best practices, not just SQL syntax.
You need to convince clients of your dbt expertise fast. A badge with a certified AI score and level is a sharper signal than a list of past clients on your portfolio site.
You learned dbt through a course or self-study and want to stand out from the crowd. The badge gives a credible signal of your real level when personal projects are hard for recruiters to assess.
You're building or evaluating a dbt team. Taking the badge yourself helps you calibrate the expected level and structure your technical interviews around the questions that actually matter.
Where and how your dbt badge will help you day to day.
You're applying for an analytics engineering role at a fast-growing startup. Instead of just listing dbt on your resume, you share your Plume badge URL showing your score and the themes evaluated. The recruiter gets an objective read on your level before the first call.
A CTO asks about your dbt expertise before handing you a Snowflake data warehouse overhaul. You send your badge report highlighting your strengths in model architecture and CI/CD, and you close the contract without an extra technical test.
You're pushing for a move from data analyst to analytics engineer at your current company. Your manager is on the fence about your dbt depth. You take the badge and share the report at your annual review — the score and per-criterion comments back your case.
A head of data has the whole analytics team take the badge to surface collective gaps in testing and dbt documentation practices. The results shape the quarterly L&D plan.
You retake the badge six months after shipping your first incremental pipeline on BigQuery. Comparing the two scores shows you exactly how much you've grown on materialization strategies and macro authoring.
You embed your dbt badge link in your LinkedIn profile and personal site. Recruiters who land on your profile see your certified level immediately, with no need to ask for a take-home assignment.
A few minutes to check you have everything you need.
At the end of your session you don't just get a score — here's everything that awaits you.
You get a numeric score and a certified level (Novice, Proficient, Advanced, or Expert) that accurately reflects your dbt command — from model architecture and incremental strategies to macros and CI/CD.
A structured report breaks down your strengths and areas for growth across all 5 evaluation criteria: dbt technical depth, concrete examples, architecture, CI/CD, and ecosystem awareness.
Your exam audio is securely stored and accessible only to you. Replay it to prep for upcoming interviews or spot formulations you want to sharpen before your next attempt.
A public URL lets you share your badge on LinkedIn, in your CV, on your portfolio, or in a recruiter email. The badge displays your score, level, and the topics covered in the exam.
Discover related skills you can validate with Plume.
A 15-min oral exam with an AI, a shareable badge for your recruiters.
Choose this badge · €19.99
