Staff AI Engineer – Business Systems

Responsibilities

AI solution architecture

• Design end-to-end agentic solutions and determine when use cases should query source systems directly versus use a unified data model.
• Partner with stakeholders to identify high-value use cases and translate requirements into controlled AI workflows.
• Select AI, conventional automation, or no new technology based on the use case.
• Create reusable architecture patterns for agents, tools, APIs, MCP servers, prompts, evaluations, and human-review workflows.
• Produce solution designs, security flows, deployment patterns, and technical standards.

AI engineering and system enablement

• Build AI agents, orchestration services, enterprise applications, and reusable platform components.
• Deliver workflows for close and reporting, procurement, forecasting, billing, and compliance monitoring.
• Establish secure, primarily read-only AI connections to approved business systems, beginning with NetSuite and extending to adjacent Finance and enterprise platforms.
• Preserve source-system authentication, authorization, user-level entitlements, rate limits, and audit trails.
• Implement citations, evidence links, deterministic checks, exception handling, and safe action boundaries.

Prototype-to-enterprise delivery

• Assess business-built prototypes for value, architecture, security, maintainability, and control readiness.
• Refactor or rebuild approved prototypes into tested, monitored, and supportable enterprise applications.
• Establish development, test, and production environments, release pipelines, incident response, and rollback controls.

AI platform strategy

• Evaluate AI models, agent frameworks, connectors, and enterprise platforms.
• Run structured proofs of concept assessing security, accuracy, integration, scalability, user experience, cost, and vendor viability.
• Maintain platform standards and recommend adoption, retention, replacement, or retirement decisions.

Organizational enablement

• Create documentation, reusable patterns, and reference architectures.
• Coach teams on agent design, prompts, evaluation practices, and safe operating boundaries.
• Establish user and process-owner feedback loops and use adoption, task success, efficiency, trust, and support signals to guide iteration.

Finance, SOX, and compliance

• Translate Finance, Security, Privacy, SOX, and SSDLC requirements into technical architecture and application controls.
• Implement least privilege, segregation of duties, logging, retention, evaluation, change control, and audit evidence.
• Require deterministic validation and reconciliation for financially material outputs.
• Support SOX walkthroughs, control testing, audits, risk assessments, and remediation while escalating formal approval to control owners.

Requirements

• 8+ years of experience in software, platform, integration, solution engineering, or enterprise applications, including meaningful hands-on production ownership in complex environments.
• Strong Python and/or TypeScript skills.
• Experience with APIs, MCP or comparable tool protocols, enterprise authentication, and distributed-system design.
• Practical experience building production AI systems using agents, tool use, retrieval, structured outputs, evaluations, and monitoring.
• Familiarity with leading LLM platforms and agent frameworks such as OpenAI, Anthropic, Gemini, LangChain, Semantic Kernel, or comparable technologies.
• Experience with prompt and context engineering.
• Strong solution-architecture judgment across security, reliability, performance, cost, observability, and supportability.
• Working knowledge of enterprise Finance processes, including general ledger, close, reporting, procure-to-pay, order-to-cash, forecasting, and management reporting.
• Working knowledge of compliance-by-design, including access, segregation of duties, change management, interfaces, automated controls, completeness and accuracy, and audit evidence.
• Ability to communicate with engineers, Finance leaders, control owners, Security, and executives.

Nice-to-haves

• Experience with ERPs, data platforms, frontier AI platforms, agent frameworks, or comparable enterprise technologies.
• Experience building internal enterprise applications.
• Hands-on experience implementing SOX controls or working in a public-company or audit-regulated environment.

Success measures

• Time from approved use case to controlled production and sustained adoption, with measurable business value.
• Reduction in manual effort and business-process cycle time; improvement in decision quality or service levels.
• Accuracy, groundedness, reconciliation success, and production reliability of deployed agents.
• User adoption, task success, stakeholder trust, and support burden for production workflows.
• Latency, operating cost, and cost per successful task.
• Reuse of approved architecture patterns and components.
• Number of viable prototypes transitioned into governed enterprise solutions.
• Security, SOX, and audit findings; evidence completeness; incident rate and remediation time.
• Quality and timeliness of AI platform evaluations and roadmap recommendations.