Staff Enterprise Architect

GitLab

16d•$168,000 - $285,600•Remote

About The Position

As a Staff Enterprise Architect, you will design and evolve how GitLab's internal systems are configured, integrate, perform and scale to support a rapidly growing business. Reporting to the Director, Enterprise Architecture and AI Intelligent Automation, you will apply all four architecture lenses (business, application, data, and technology) to the systems in your scope, treating AI as a standing consideration inside each rather than a separate workstream. Enterprise-wide strategy, portfolio direction, and standards ownership sit with the Director. Your job is depth: turning that direction into designs detailed enough to build, and being the person trusted with the hardest technical problems in the enterprise. You will not manage people. You will lead through architecture, as the partner to our Applications engineering teams and to Integration, RPA, and Intelligent Automation, and as a reviewer on the Architecture Review Board, where the aim is to make architecture an enabler rather than a gate.

Requirements

Deep expertise in application architecture and data architecture, with working fluency in business process and technology architecture, AI as a working lens across all four, and the judgment to know which lens a given problem needs.
Strong hands-on background in integration, automation, and applied AI, including retrieval-augmented generation, tool calling and agent orchestration on a platform such as Claude, plus the evaluation and guardrails that make them safe to deploy, the judgment to know where AI belongs and where it does not, and fluency using AI in your own daily work, including AI-assisted development, rather than only designing it for others.
Hands-on experience architecting core business systems, including ERP, CRM, CPQ, billing, financial, and customer service platforms, with the discipline to keep them standard, upgradeable, and SOX-compliant, plus fluency in the processes that cross them, from quote-to-cash through record-to-report.
Expert-level knowledge of data modeling, integration architecture patterns, and API design for enterprise systems, plus practical experience with data warehouse technologies and ELT, ETL and reverse ETL pipelines, across operational and analytical use cases.
Depth in iPaaS tooling and the operational realities of running integrations in production, including cost and consumption optimization, plus working knowledge of the cloud platforms those integrations run on.
A track record of driving architectural change through influence rather than authority: designs that survived implementation, standards adopted on platforms owned by other teams, a clear technical position held in review forums without losing the partnership, and trade-offs explained in plain business language to Director and VP stakeholders, in writing and async.

Responsibilities

Map the current-state processes and capabilities supported by the systems in your scope, tying each capability to its supporting systems, owners, and systems of record, and identify and resolve where those systems create friction, manual handoffs, or duplicated work, including handoffs an AI agent could remove entirely.
Architect end-to-end processes that span front-office and back-office systems, including new product introduction, CPQ and quote-to-cash, billing and revenue recognition, renewals, and record-to-report, together with the customer service processes that run alongside them, including case and escalation management, entitlements, and service level commitments, rather than optimizing any one system in isolation.
Redesign workflows before automating them, so we aren't encoding broken processes into new tooling. For each step, decide what an AI agent should handle, what deterministic automation should handle, and what stays with a person.
Serve as an embedded architect for Applications engineering teams, producing high level designs and patterns for changes to Salesforce, NetSuite, Zuora, Zendesk, and other core platforms. Favor configuration and supported extension patterns over customization, evaluate each platform's native AI before anything bespoke is built, and keep system roadmaps aligned with enterprise standards.
Design for maintainability and supportability: clear service ownership, observability and error handling in integrations, documented runbooks, and realistic total cost of ownership.
Go deep on the platforms in your scope, staying current on release notes, published limits, API behavior, and known constraints, so designs account for what the platform will actually do rather than what the documentation promises.
Govern application change by reviewing significant build proposals, data model changes, and third-party package additions before they are committed, giving teams a clear and timely decision, and keeping platforms on supported versions and ready to upgrade.
Run quarterly health checks on Salesforce, NetSuite, Zuora, Zendesk, and Workato, measuring customization footprint against each platform's published ceilings for API and governor limits, automation, storage, transaction and ticket volumes, AI consumption, and licenses. Publish a report card showing headroom, what is trending toward a limit, and who owns the fix.
Architect data flows across operational and analytical planes, including the retrieval and grounding patterns that make enterprise data usable by AI, alongside API-first integrations, event-driven patterns, batch pipelines, data warehouse models, and reverse ETL back into business systems.
Design data access and handling patterns that meet privacy, residency, and SOX requirements, setting explicit boundaries on what AI models and agents may read, retain, or act on, plus field-level controls for sensitive customer and financial data.
Establish integration and data contracts between systems, so downstream consumers aren't adversely affected by upstream schema changes.
Define platform and integration patterns for the systems in your scope, spanning iPaaS (Workato), identity and access management, secrets handling, and the cloud services those systems depend on. Define how each system exposes its capability to AI agents, through APIs, tool interfaces, and scoped permissions.
Design for reliability and scale, covering environment strategy, sandbox and test data management, and non-functional requirements, while CI/CD and disaster recovery remain theirs to own.
Design authentication, authorization, and service-to-service access for the systems in your scope, applying the standards set by Security and the Director.
Recommend, use case by use case, whether a need is best served by domain-specific AI embedded in a business application or by a shared enterprise AI capability, applying the enterprise AI team's decision and model governance framework, and assemble the evidence each recommendation rests on: data access and sensitivity, reuse across functions, and cost.
Monitor and optimize cloud and platform consumption, covering right-sizing, license and API-limit management, and eliminating spend on unused or duplicated capacity.
Produce the designs teams build from: target-state architectures, decision records, reference patterns, and high level designs for major projects and significant platform changes. Specify them in enough detail to build against: data models, field-level mappings, error and retry behavior, failure modes, and how AI output is evaluated where a design uses it. Every design must satisfy SOX, internal control, and security requirements.
Co-author designs with engineering teams and stay with them through build, making the call on trade-offs, resolving ambiguity in the spec, and unblocking decisions so delivery doesn't stall.
Take the hardest technical problems in your domains personally, building proofs of concept when a decision cannot be settled on paper, including the production issues no one else has been able to root-cause.
Collaborate very closely with adjacent technology teams, including Data, Security, and Infrastructure, so that architecture is thought through end to end and serves the whole enterprise rather than one function at a time.
Maintain a prioritized inventory of technical debt across your domains and drive it down by attaching remediation to funded projects, upgrades, and migrations already touching that surface. Root-cause production defects rather than clearing them, fixing the design so the same class of defect stops recurring.