AI Senior QC
SoftServeAbout The Role
SoftServe is delivering the AI Student Tutor education company — a Canvas-integrated, context-aware AI tutor and learning coach. The product routes student conversations through a single-assistant orchestration layer to specialized AI capabilities (tutoring, content generation, safety/distress handling), and must meet strict pedagogical and safety data requirements ahead of a production pilot launch.
The Senior QC Engineer owns quality across this engagement, working within the SoftServe delivery team alongside AI/ML engineering, application engineering, product, and data engineering, and coordinating directly with client's Product, SME, IT, and Security/Compliance stakeholders through UAT and pilot-readiness activities.
Responsibilities
SoftServe is delivering the AI Student Tutor education company — a Canvas-integrated, context-aware AI tutor and learning coach. The product routes student conversations through a single-assistant orchestration layer to specialized AI capabilities (tutoring, content generation, safety/distress handling), and must meet strict pedagogical and safety data requirements ahead of a production pilot launch.
The Senior QC Engineer owns quality across this engagement, working within the SoftServe delivery team alongside AI/ML engineering, application engineering, product, and data engineering, and coordinating directly with client's Product, SME, IT, and Security/Compliance stakeholders through UAT and pilot-readiness activities.
Responsibilities
- Own creation of the Acceptance Testing Plan (ATP), defining acceptance criteria, test types, roles and responsibilities, test data, evidence expectations, defect severity definitions, retest process, and formal sign-off requirements, in coordination with Covista.
- Design and execute the overall test strategy spanning functional, integration, regression, load/performance, accessibility, safety, and pedagogical-quality testing.
- Test the AI agent application itself, including single-assistant routing/orchestration logic, prompt governance and version-tracking, persona-adaptive behavior, content-validation and grounding/citation pipelines, and resistance to identity-override and prompt-injection attempts.
- Run safety scenario testing across low, medium, high, and critical distress-handling cases, measuring detection accuracy and false-negative/false-positive rates.
- Drive UAT: build test scripts, execute with client's Product, SMEs, IT, Security/Compliance, and pilot stakeholders; manage defect triage, disposition, and re-test.
- Serve as the primary client-facing point of contact for quality matters: communicate testing status, defect trends, and pilot-readiness risk directly to client's Product, SMEs, and other UAT stakeholders.
- Jointly verify and reproduce reported bugs with Covista during UAT and Hypercare, confirming severity, scope, and expected vs. actual behavior before logging.
- Translate verified client-reported issues into clear, actionable defect write-ups (repro steps, environment, evidence, severity) and hand them off to the Europe-based delivery team, working within the daily US/Europe overlap window to ensure timely triage and unblock next-day resolution.
- Proven experience leading QA/QC for complex, multi-workstream software programs, including authoring test strategies and acceptance test plans from scratch.
- Experience testing AI/LLM-based or conversational agent systems — prompt-driven behavior, RAG/grounding validation, model-output quality rubrics, and safety/guardrail testing.
- Familiarity with LMS integration testing (Canvas/LTI, SSO, roster sync) is a plus.
- Strong written documentation skills for test plans, traceability matrices, defect reports, and readiness packages.
- Strong client-facing communication skills, comfortable running bug-verification calls/sessions directly with client stakeholders and writing defect descriptions clear enough for a remote team to act on without follow-up questions.
- Availability during the US–Europe daily overlap window to support live client bug verification and same-day handoff to the Europe-based team.