Country or region and language

    Choose your country or region and your language.

    Suggested for you

      All countries and regions

        Choose your character

        It shows in the top bar and on your profile. Change it any time.

        Interviews in Dagster

        Work in the tool while you explain your decisions.

        Choose a role and level. Each interview has a new task, follow-up questions and a report that checks your work and your explanation.

        Speak or type. Sign in to save your results across devices.

        Timed assessment

        • 6 practical questions
        • 35 minutes
        • Mid level

        Questions are drawn from a bank of 18. Your report includes a skill breakdown and lessons to practice next.

        Practice interviews

        Junior tasks cover the fundamentals. Mid and senior tasks add debugging, changing requirements and review decisions.

        6 interviews

        • Data engineer
        • Junior
        • 22 min

        A signed delivery liability

        Birch Delivery recognizes only signed delivery rows. Build source_rows, accepted and liability assets. Filter status equal to signed in accepted, sum amount in liability, and preserve both declared dependencies. Negative signed adjustments remain eligible. Read hub_case.py for the supplied rows.

        Start interview
        Skills and rubricAsset dependencies
        • I pass source_rows into accepted and accepted into liability because the asset graph must describe the data the total uses. I retain signed negative adjustments and reconcile the sum to the eligible rows. File order does not establish a dependency.
        • Complete the requested work and identify the source fact that makes it valid.
        • I compare a signed negative adjustment with an unsigned positive delivery because filtering only positive amounts would overstate the liability. I also inspect the declared upstream keys rather than infer the graph from a correct total.
        • Analytics engineer
        • Junior
        • 22 min

        A dispatch partition with a correction

        Clover Dispatch needs a daily dispatches asset, partitioned from November 1, 2026. Use context.partition_key to retain only that day from hub_case.ROWS. The check materializes November 5 and November 6 separately; neither partition may contain the other day.

        Start interview
        Skills and rubricDaily partitions
        • I declare DailyPartitionsDefinition and filter each row by context.partition_key because a partition label alone does not restrict the source. I compare the stored rows of both dates and retain the correction with its own source date.
        • Complete the requested work and identify the source fact that makes it valid.
        • I check both dates because hard-coding November 5 can pass its visible materialization and fail the next partition. A late correction must update the selected historical partition instead of appearing in the wall-clock day.
        • Data engineer
        • Mid
        • 30 min

        A repeated receipt identifier

        Copper Receipts receives repeated receipt IDs. Add unique_receipts on receipts, return an AssetCheckResult that passes exactly when every ID is unique, and record the number of extra rows as duplicates metadata. Register the check with Definitions. Empty input is valid.

        Start interview
        Skills and rubricAsset checks
        • I test an ID repeated three times because two extra rows differ from one duplicated identity. I also test an empty list so the check returns a valid result instead of raising an exception.
        • I make the check blocking with the required error severity and test that downstream publication does not execute because an ordinary failed check can report a problem without enforcing that publication boundary.
        • Use a counterexample and a readback to defend the result under changed inputs.
        • Analytics engineer
        • Mid
        • 30 min

        Rows in a published manifest

        Hawthorn Freight publishes an accepted_manifest asset containing only rows with status signed. Return those rows and attach row_count metadata equal to the returned row count. The metadata must count accepted rows, not every source row.

        Start interview
        Skills and rubricMaterialization metadata
        • I include unsigned rows in the fixture because counting the source would otherwise look correct. A successful materialization with a plausible row count still needs a source-to-output reconciliation.
        • I define the deduplication key and correction precedence before changing the output because a distinct-ID count does not decide which repeated row to retain. I rename or document the metadata definition and test repeated IDs with conflicting values.
        • Use a counterexample and a readback to defend the result under changed inputs.
        • Data engineer
        • Senior
        • 38 min

        A transient export on its third attempt

        Larch Exports fails its first two calls with a transient ConnectionError, then returns the supplied rows. Add an asset RetryPolicy with two retries to exported_rows and return hub_case.fetch(). The check requires a successful third attempt without swallowing the exception.

        Start interview
        Skills and rubricRetries
        • I classify the rejected credential as a permanent failure for this run and route the repair to its owner because repeating the same unauthorized request does not restore access. I retain the failed run and avoid publishing substitute empty data.
        • I retain the original request, input identities, code or commands, checked output and run evidence because another engineer needs a reproducible result. I name the remaining assumption and the next check that could change it. The browser model cannot establish production performance or external system behavior.
        • Define the replay or recovery boundary and the evidence required before publication.
        • Data engineer
        • Senior
        • 38 min

        A redelivered file notification

        Willow Landing lists immutable file version IDs in hub_case.FILES. Add incoming_files targeting loaded_marker. Return one RunRequest per ID with that exact ID as run_key. Evaluate two ticks over the unchanged list; each version may launch once across both ticks.

        Start interview
        Skills and rubricSensor run keys
        • I give the corrected version a new run key while keeping exact-version redelivery deduplicated because a path can identify several different payloads. I rebuild affected outputs under an explicit correction policy rather than append the correction blindly.
        • I retain the original request, input identities, code or commands, checked output and run evidence because another engineer needs a reproducible result. I name the remaining assumption and the next check that could change it. The browser model cannot establish production performance or external system behavior.
        • Define the replay or recovery boundary and the evidence required before publication.

        Interview Hub access

        The Interview Hub is included in All Access. Choose a plan or sign in to your account to continue.

        View plansSign in or join

        Your Work Gym account

        Save your progress and pick up on another device. Your browser's progress comes with you.

        or

        You can keep practicing without an account.

        Change your plan

        Checking the price

        Help and feedback

        Ask a question, tell us what is not working, or share an idea. We reply by email.

        What is it about?

        We use it only to reply to you.