Must be located near or within the area of Cincinnati, Ohio.
Toshiba Global Commerce Solutions is seeking a Director, Software Engineering: Retail Platform Delivery (ELERA) to own end-to-end delivery of the ELERA retail platform across our strategic global retail partners, including Fortune 100 grocery, warehouse club, and general merchandise accounts.
This role collapses Development, Test, and Professional Services Engineering into a single accountable function and operates as the customer-facing engineering authority for strategic programs. You own what ships, how it scales, how it recovers, and how quickly the team learns.
This role runs on Toshiba’s agentic AI development ecosystem as the default way of working, not a side experiment. Agents participate in requirement decomposition, code generation, test authoring, PR review, defect triage, and release scoring. You are expected to measure and compound that leverage.
What You Own
• Delivery velocity AND production reliability (DORA metrics with published targets)
• Unified engineering and quality accountability in one org
• Customer-facing technical authority for strategic retail accounts
• AI-native development workflows that compound team output quarter over quarter
• Production outcomes: what ships, how it performs, how it recovers, what it costs to run
This is not an engineering-only or QA-only role. You own the outcome, the telemetry that proves it, and the workflows that compound it.
Key Responsibilities
Customer-Facing Engineering Leadership
• Primary engineering authority for customer programs spanning POS, payments, loyalty, and store systems
• Partner with customer leaders to define success criteria, operational acceptance, rollout strategy, and risk posture
• Translate real-world retail workflows into architecture decisions, API contracts, and system behavior
• Maintain architecture decision records (ADRs) and a visible technical debt register tied to delivery plans
End-to-End Delivery Ownership (DORA-Native)
Own the four DORA metrics as published, measurable targets:
• Deployment Frequency: on-demand for services, scheduled-safe for store edge
• Lead Time for Changes: under 24 hours for services
• Change Failure Rate: under 10% trending down
• MTTR: under 1 hour for customer-impacting incidents
• AI-built feature branches with pre-merge release testing: agents build the full feature on an isolated branch; release-grade testing (functional, performance, security, integration) runs on that branch and must pass before merge. Main stays releasable at all times.
• Progressive delivery is mandatory: feature flags, canary, blue/green, shadow traffic. Decouple deploy from release.
• Zero scheduled-downtime windows for cloud services; store edge deploys are non-disruptive to live lanes
Unified Engineering + Quality Model
• Test-Driven Development is the baseline. Tests exist before code. No exceptions for new code paths.
• Quality gates block merges: coverage thresholds, mutation score, static analysis, SAST, dependency and license scanning, SBOM generation
• Contract testing (consumer-driven) between every service boundary; OpenAPI and AsyncAPI versioned and governed
• Chaos engineering in staging, quarterly game days, production-like load testing ahead of peak retail events
• Shift-right validation: synthetic monitoring, canary analysis, error budget enforcement
• Own metrics: defect leakage, change failure rate, cycle time, escape rate, production incident rate
AI-Native Development (Agentic Delivery)
Leverage Toshiba’s agentic development ecosystem as the default delivery model:
• AI-assisted engineering: requirement decomposition, design-to-code, test generation, refactoring at scale, PR review
• Agent-driven workflows: defect triage and clustering, root cause analysis from telemetry, release readiness scoring, customer-reported incident classification
• MCP-integrated tooling: every engineer works with context-aware agents wired to source, tickets and telemetry
• Agents as team members: RaaS (RAG-as-a-Service), Sentinel AI for test authoring and execution, Agentic TPM for program management, CloudOps Sandbox for environment provisioning, ELERA Agent Registry for discovery
• Measured leverage: agent-authored code accepted, test cases generated, incidents auto-triaged, AI-equivalent headcount by role tracked quarterly
• Feedback loops: production telemetry feeds AI-ranked backlog and refactor targets
Scalability, Performance, and Reliability (SRE Discipline)
• SLOs and error budgets defined for every customer-facing service; error budget exhaustion blocks feature work
• Observability-as-code: OpenTelemetry instrumentation is required before merge; service.version tagged; traces, metrics, and logs correlated by default
• Performance engineering: load, stress, endurance, and soak testing sized for peak retail events (Black Friday, holiday, payroll cycles)
• Edge resilience: stores survive network partitions; offline transaction handling; store-and-forward retry with exponential backoff and jitter; blue/green at the edge
• Horizontal scale to tens of thousands of lanes across multi-tenant, multi-region deployments
• Capacity planning tied to FinOps: cost per transaction tracked and optimized, not assumed
Platform Engineering & Developer Experience
• Treat the engineering org as the customer of an internal developer platform (AI Dev Portal as the front door)
• Golden paths: paved roads for new services, standard observability, standard security posture, standard CI/CD
• Self-service environments: CloudOps Sandbox, on-demand synthetic test data, ephemeral preview environments per PR
• Developer productivity instrumented: DORA plus SPACE metrics; time-to-first-commit for new hires; PR review latency; local build time
Professional Services Engineering Model
• Customer-specific extensions delivered without fragmenting the platform
• Config-driven first, code-driven by exception. Every bespoke extension has a sunset plan or a promotion path to core.
• Reusable solution accelerators for common retail integrations (payments, peripherals, back office, tax, loyalty, CRM)
• Integration contracts versioned and governed; breaking changes require deprecation windows and migration tooling
• Clear engineering handoff to customer operations and support, with runbooks, dashboards, and on-call rotations live at go-live
Release Governance and Operational Readiness
Engineering-driven release gates, not process theater:
• Functional completeness, validated by automated acceptance
• Performance validation against published SLOs
• Integration stability, with contract tests green
• Observability live (dashboards, alerts, runbooks) before traffic
• Security posture verified (SAST, DAST, SCA, SBOM, secrets scanning, threat model reviewed)
• Rollback tested, not assumed
• Post-release: automated canary analysis, 24-hour watch, blameless incident reviews with 48-hour turnaround and action-items tracked to closure
Security and Compliance Engineering
• Shift-left security: threat modeling at design time, SAST in every PR, DAST in staging
• Compliance built into the pipeline: PCI DSS, SOX, GDPR controls are code, not audit artifacts
• Software supply chain: SBOM generation, signed artifacts, SLSA build-level targets
• Secrets management, zero-trust service-to-service authentication, least-privilege IAM
Leadership and Organizational Impact
• Lead cross-functional teams spanning Engineering, embedded Quality, and Professional Services
Develop engineering leaders who:
• Own outcomes, not tasks
• Think in systems and data, not features
• Hold the room at the customer executive table
Drive a culture of:
• Accountability without blame
• Evidence-based decisions (telemetry, metrics, experiments)
• Continuous improvement with measurable deltas
What Good Looks Like
• DORA elite: daily deploys, hours-to-production, CFR under 10%, MTTR under an hour
