Toshiba Global Commerce Solutions is seeking a hands-on Principal Software Engineer to own the speed, determinism, reliability, and evolution of a large browser-driven BDD execution platform used to validate complex retail point-of-sale behavior.
This is not a manual-QA lead, test-script maintenance, or paper-architecture role. You will work directly in Java runner and lifecycle code, Cucumber features and steps, the shared POS UI abstraction, Selenium WebDriver and Grid, Docker execution, CI scheduling, OpenTelemetry traces, and failing scenarios. You will reproduce difficult failures, implement the fixes, and prove them through repeated execution.
You will also oversee agentic test-engineering workers. You will bound their work, provide feature, code, trace, and performance context, review and correct output, block shortcuts such as sleeps or retry masking, and remain accountable for every accepted test or platform change.
Your Mission
Turn browser-driven BDD into a production-grade execution platform that is fast, deterministic, diagnosable, and trusted as a release signal. Begin with the current Selenium-centered estate, harden it, and guide any Selenium/Playwright evolution through representative evidence rather than preference.
Technical Context
- Java and JavaScript applications, Maven, Docker-based local and CI environments, Kubernetes and Helm assets, and GitLab/Jenkins workflows.
- A large Cucumber suite spanning fixed-lane, specialty, and base POS flows through Java step definitions and a shared UI abstraction.
- Selenium Grid, parallel workers, seeded test data, service and device dependencies, reruns, quarantine/re-enablement, priority tags, per-scenario and per-step OpenTelemetry, and an emerging Playwright-compatible path.
What You Will Own
- Work directly in the runner, lifecycle hooks, step definitions, shared UI abstraction, browser infrastructure, test data, CI configuration, and observability code that determine execution behavior.
- Own BDD platform architecture and governance across feature design, runners, domain actions, drivers, fixtures, data, reporting, tagging, compatibility, ownership, and lifecycle rules.
- Harden Selenium/WebDriver/Grid behavior, including asynchronous application state, frame switching, browser/session lifecycle, stable locators, meaningful assertions, condition-based waits, cleanup, and failure recovery.
- Eliminate shared state, test-order dependence, data collisions, incomplete teardown, and environment contamination through deterministic setup, isolation, idempotency, and cleanup.
- Profile the full execution path, from discovery and setup through each UI step, teardown, evidence collection, queueing, scheduling, and infrastructure wait; improve parallel efficiency and long-tail behavior without saturating shared dependencies.
- Classify and eliminate flake by root cause; govern reruns and quarantine as transparent, temporary containment with ownership and exit criteria, not as proof of reliability.
- Extend browser-to-backend evidence and OpenTelemetry so the first failure contains enough correlated data to begin diagnosis without a manual rerun.
- Define risk-based CI policy for blocking, asynchronous, quarantined, and informational coverage while preserving visible release confidence.
- Direct agentic workers for test-gap analysis, scenario generation, failure clustering, flake investigation, performance analysis, refactoring, and Selenium/Playwright experiments, with compilation, targeted execution, repeated-run stability, and regression evidence required before acceptance.
First-Year Success Outcomes
- A published baseline captures scenario and step duration, queueing, scheduling delay, long-tail contributors, parallel efficiency, rerun behavior, true flake, infrastructure failures, quarantine age, and diagnosis time.
- Critical-path and full-suite execution improve measurably from baseline through code, lifecycle, scheduling, and capacity changes while preserving intended risk coverage.
- Recurring causes involving lifecycle races, frame handling, state leakage, cleanup, waits, locator instability, or Grid saturation are eliminated or materially reduced with repeatable evidence.
- Failed scenarios produce correlated browser, runner, environment, test-data, and backend evidence that makes first-pass diagnosis practical.
- A Selenium/Playwright direction is documented from representative benchmarks and maintainability evidence, without assuming migration is the answer.
- Agentic test-engineering workflows improve accepted output, stability, and diagnosis time without increasing weak assertions, retries, hidden quarantine, or test volume without value.
What You Bring
- Substantial software engineering, quality engineering, or test-platform experience with Staff or Principal-level influence and a recent hands-on record of changing framework or platform code in real execution environments.
- Deep hands-on expertise with Java or another JVM language, Cucumber/Gherkin, Selenium WebDriver, and Selenium Grid or comparable distributed browser infrastructure.
- Ownership of a large browser-driven BDD or end-to-end test estate with shared framework components, parallel execution, and multi-team adoption.
- Strong understanding of browser and test lifecycle design, asynchronous behavior, frames, synchronization, locators, assertions, state isolation, setup/cleanup, and deterministic test data.
- Experience profiling and improving execution performance, sharding, scheduling, queueing, tail latency, capacity, and infrastructure utilization.
- Strong experience with Maven, GitLab/Jenkins or equivalent CI platforms, Docker, Kubernetes, Helm, environment orchestration, and pipeline-as-code.
- Practical observability expertise using OpenTelemetry or comparable traces, metrics, logs, dashboards, and correlated failure evidence.
- Experience directing AI coding or test-engineering agents through bounded tasks, context design, output review, repeated-run validation, and prevention of automation shortcuts.
Preferred Qualifications
- Playwright, WebDriver BiDi, or evidence-based browser-framework migration or coexistence experience.
- Retail POS, payments, store devices, peripherals, stateful transaction flows, or intermittently connected edge environments.
- Statistical flake analysis, queueing or scheduling optimization, service virtualization, seeded data, and Docker-orchestrated dependencies.
Why This Is a Principal Role
The Principal level reflects the need to operate across business behavior, framework code, browser protocols, distributed infrastructure, data, dependent services, CI policy, observability, and release confidence without stepping away from implementation. The Developer Experience Platform Principal owns the broader developer journey; this role owns the BDD/browser execution platform itself.
We offer:
- Hired directly by Toshiba and 100% under payroll.
- Benefits of the law (IMSS, INFONAVIT, December day voucher, vacation voucher)
- Life insurance, major medical expenses with dental and visual coverage. (For the employee and direct family members).
- Minor medical expenses insurance.
- Savings fund for employees, capped at the limit of the law.
- Grocery coupons.
- Pension plan.
- 12 days of vacation plus 6 floating days.
EO: Toshiba Global Commerce Solutions is an equal opportunity/affirmative action employer that evaluates qualified applicants without regard to age, ancestry, color, religious creed, disability, marital status, medical condition, genetic information, military or veteran status, national origin, race, sex, gender, gender identity, gender expression and sexual orientation or any other protected factor. We also consider qualified applicants regardless of criminal histories, consistent with legal requirements.