QA Engineer – Creative Automation (100% Remote) (LATAM ONLY)

Allshore Talent•Ambinany, B

4d•Remote

About The Position

AllShore Talent is seeking a mid-to-senior QA Engineer specializing in creative automation to validate end-to-end creative automation and digital media production workflows. The work spans Airtable intake and status control, AWS orchestration, Adobe Photoshop APIs, and Frame.io review and approval. The current use case focuses on static banners and Photoshop outputs, with motion graphics and video workflows in the pipeline, including one moving into implementation. The role calls for a methodical approach to tracing distributed workflows, finding edge cases, maintaining state integrity, and verifying data isolation. Relevant backgrounds include digital media workflow engineering, ad-tech production pipelines, creative automation, and cloud-native integration QA.

Requirements

Mid-to-senior QA engineering experience aligned with digital media workflow engineering, ad-tech production pipelines, creative automation, or cloud-native integration QA.
Strong end-to-end troubleshooting, edge-case testing, state integrity, and data isolation skills.
Practical understanding of Airtable field types, formulas, automations, and trigger actions; ability to validate Action = Execute and Automation Status transitions from Pending to Queued to In Progress to Review, Approved, or Finalizing.
Experience checking Airtable field hygiene, null and fallback values such as NA, Automation Version Stack ID tracking, and Invalidate Asset Cache and Creative Allocation overrides.
Strong practical knowledge of Frame.io as an automated review and approval surface, including version stacking, Approved asset status hooks, share-link generation, and structural validation.
Experience with Frame.io API authentication nuances, including server-to-server (S2S) authentication and workarounds, custom headers, and rate-limiting behavior.
Solid conceptual understanding of Adobe Photoshop artboards, smart objects, layer hierarchies and masks, text layers, typography leading and sizing, and bounds calculation.
Familiarity with Adobe Photoshop APIs and smart-object replacement mechanics; ability to inspect and validate PSD/PSB, JPG, PNG, and GIF outputs, including raster and animated renditions as applicable.
Comfort navigating AWS SQS queues, DynamoDB locking and slot tracking, Step Functions/state machines, Lambda, and S3 asset storage and overwrite behavior.
Ability to trace execution logs, correlation IDs, and failure payloads to distinguish upstream intake errors, service timeouts, and circuit breaker trip states.
Experience testing asynchronous and event-driven systems, webhook de-duplication, idempotency, concurrency, throttling, resilience, circuit breakers, layered caching and invalidation, and asset and folder synchronization edge cases.
Understanding of bounding-box and focal-point detection, native cropping, Firefly Generative Expand, and mask boundaries; keen eye for digital display production and design standards across multiple output formats and aspect ratios.
Deep comfort working in Jira Cloud with Epics, Stories, Bugs, acceptance criteria, and detailed, reproducible defect reports supported by assets and logs.

Nice To Haves

Familiarity with Adobe Premiere Pro, After Effects, Motion Graphics templates (MoGRTs), and Adobe Media Encoder, particularly presets, video codecs, output formats, and resolutions, to support upcoming motion graphics and video use cases.

Responsibilities

Validate asynchronous, event-driven pipelines triggered by intake actions, tracing workflows from Airtable through AWS and Photoshop APIs to Frame.io.
Verify idempotency, de-duplication, race-condition handling, and queue throttling, including duplicate Frame.io Approved webhooks.
Test recovery from network interruptions, expired tokens, corrupt or oversized source media, missing manifest keys, and simulated API service limits; verify circuit breaker thresholds, queue pausing, slot release, and manual or administrator recovery.
Design deterministic cache lifecycle tests that distinguish reusable asset-level caches (Creative/DMC) from isolated template-level manifests; confirm timestamp changes and asset re-uploads invalidate caches and prevent stale file delivery.
Reconcile assets, folders, and version stacks; test URL-encoded filenames, duplicate same-name folder structures, version stack collisions, and protection of active in-progress uploads from deletion.
Validate AI cropping and graphical transformations across human and non-human subjects in landscape and portrait source files; assess bounding-box and focal-point detection, evaluate use of Firefly Generative Expand versus native cropping, and verify masks preserve native boundaries.
Inspect typography, font substitutions, leading, multi-line wrapping, bold ranges, safe-zone margins, CTA button bounds, disclaimer visibility, and ratings bug placement across aspect ratios.
Document reproducible, step-level defects in Jira Cloud with complete asset and log context, and evaluate results against acceptance criteria associated with Epics, Stories, and Bugs.