QA + AI Engineer

Mandatory Skills

Summary
ALP-U is hiring a contract QA engineer to embed in the team as a full sprint member. The role provides a quality pass on completed work before promotion to the internal Staging environment, owns the staging release train, and maintains the health of the automated test suites.




The engineer participates in standups, planning, and retrospectives and is accountable to the same sprint commitments as the rest of the team. The role does not relieve engineers of responsibility for testing and validating their own work; developer ownership of feature code and its tests is retained.




The engineer works to the team's codified QA standard (the qa-start skill and its referenced conventions), applying that definition of quality rather than an external one.




Responsibilities" Release quality pass

Verify completed work against acceptance criteria, design intent, and expected product behavior before promotion to Staging.
Test on the development environment (lp-dev.taara.company), covering both the change under review and adjacent functionality at risk of regression.
File defects in Linear with reproduction steps, expected versus actual behavior, environment, and severity.
Provide Product with a concise, pre-vetted summary of each release candidate for sign-off.

Staging release train
Own the staging release train, replacing the current round-robin among engineers.
Determine which completed work is ready to promote and run the release through the team's release automation.
Promote work that passes; do not repair work that fails. Failing items return to the owning engineer with a reproduction.

Test suite health
Operate and maintain the automated Playwright, unit, and integration suites: run them, triage failures, and separate genuine defects from flaky tests.
Diagnose and repair flaky tests.
Validate test plans against code; detect duplication, dead tests, and coverage gaps; keep plan status and coverage summaries accurate.
Extend or repair specs where it is the fastest path to coverage.
Feature-test authorship remains with the engineer who builds the feature. A feature is not complete until its developer-authored tests exist and pass. The engineer does not become author of record for feature tests, and does not own architecture, deployment mechanics, or code review.
Scope
Coverage spans both Taara Base and Taara Link Planner.
Base is integration-dependent (ServiceNow, HubSpot, Sanity, GCS, and planned Prism sync); testing includes verification across these systems.
Role-based behavior must be validated across the products' permission models (Owner / Admin / Installer; Planner / Admin / Owner).
Base includes mobile-responsive surfaces for field use; Link Planner is desktop-first.
QA plans track user-facing scenarios. Infrastructure unit tests (metrics, logging, analytics, queue internals) are owned by developers and are out of scope for QA plans.
Security hardening and penetration-test remediation are out of scope for this engagement.

Required Qualifications
Demonstrated product judgment: able to derive intended behavior from a ticket and identify gaps between a passing implementation and the intended outcome.
Risk-based test prioritization.
Proficiency with AI tooling (Claude, Claude Code, or equivalent) integrated into a working QA process.
Proficiency with Playwright: able to run, interpret, triage, and repair end-to-end specs, including diagnosing flaky tests. Authoring the full suite is not required; engineers write their own tests.
Working knowledge of the stack: Next.js, React, TypeScript, browser developer tools, and application logs.
Clear, actionable written communication for defect reports and release summaries.
Ability to operate autonomously within a defined goal.

Preferred Qualifications
Prior ownership of a release or staging-promotion process.
CI/CD experience (GitHub Actions); cross-browser test debugging (Firefox, WebKit).
Experience with geospatial or mapping interfaces, data visualization, or scientific-modeling products.
Familiarity with product analytics tooling (e.g., PostHog) for instrumentation verification.

Working Model
Embedded member of the ALP-U sprint team.
Participation in standups, sprint planning, and retrospectives.
Defect tracking and release coordination conducted in Linear.
QA process governed by the team's qa-start skill and release automation