Sr AI Manager

What matters most

• Production-ready AI delivery

  The role leads hands-on design and implementation of scalable enterprise AI systems and establishes MLOps, LLMOps, and cloud DevSecOps practices. Preparation suggestion: have an end-to-end example that explains architecture, deployment, performance, and operational ownership.

• Measurable value across enterprise systems

  The posting requires multi-system AI and automation solutions, including integrations with platforms such as SAP, SFDC, and HCM, that deliver measurable business value. Preparation suggestion: be ready to explain how an integration changed a business outcome, not just how it worked technically.

• Responsible speed to delivery

  The manager is expected to accelerate iteration while guiding the Corporate AI Risk Council on the balance between adoption speed, risk, and compliance. Preparation suggestion: describe decision points and controls that allow an initiative to move quickly without bypassing review.

• Evaluation and continuous improvement

  The role calls for an evaluation- and observability-first approach to Gen AI reliability and a roadmap for continuous-learning systems. Preparation suggestion: explain how evidence from production use would inform monitoring, evaluation, and subsequent changes.

• Cross-functional technical leadership

  The manager partners with business, technology, architecture, and security teams, develops classification-aware adoption guidance, and translates AI advances into enterprise strategy. Preparation suggestion: illustrate how technical choices can be communicated as practical decisions for different stakeholders.

Likely challenges to think through

• Iteration versus governance

  Possible challenge, not a confirmed employer problem: rapid Gen AI experimentation may conflict with classification-aware data handling and risk review. Prepare a concrete approach to staging releases, applying safeguards, and deciding when an initiative is ready to advance.

• Reuse versus functional fit

  Possible challenge, not a confirmed employer problem: solutions spanning supply chain, finance, commercial, and other functions may need common architecture while accommodating different processes and enterprise integrations. Prepare to discuss where to standardize and where to adapt, using measurable outcomes to justify the trade-off.

• Learning versus reliability

  Possible challenge, not a confirmed employer problem: improving AI systems from operational feedback can introduce changes that affect reliability. Prepare to explain how evaluation, observability, and controlled updates would help distinguish genuine improvement from regression.

Topics to prepare

• Enterprise AI architecture

  The role leads scalable, end-to-end AI and automation systems that integrate multiple enterprise applications. Revise by comparing architectures for one supply-chain and one finance use case, explaining system boundaries, performance trade-offs and how each would produce measurable business value.

• Generative AI, classical AI and automation

  The posting covers all three approaches across enterprise functions. Revise by taking three hypothetical business requests and explaining when each approach fits, what data it needs and why a simpler automation might be preferable to an AI model.

• Production operations for AI

  MLOps, LLMOps and Cloud DevSecOps are explicit delivery responsibilities. Revise the concepts by tracing a solution from development through release, monitoring, rollback and updates; identify where security and reliability checks belong.

• Evaluation, observability and continuous learning

  The role calls for an evaluation-and-observability-first approach and a roadmap for continuous learning. Revise by defining quality, reliability and business-impact measures for a hypothetical AI service, then explaining what monitored results would trigger investigation or an update.

• Responsible and classification-aware AI

  The manager develops adoption guidance and advises the Corporate AI Risk Council on risk, compliance and speed. Revise by working through a hypothetical sensitive-data use case: classify the inputs, identify risks and explain which controls or review decisions would affect deployment.

Skills to revise

• Design a multi-system integration

  Experience integrating with platforms such as SAP, SFDC and HCM is required. Practice drawing an end-to-end data and control flow between a hypothetical AI service and two of those systems, including authentication, failure handling and a measurable outcome.

• Package a production-ready solution

  The posting requires hands-on delivery and proficiency with multiple AI/ML and automation tools; Azure AI/Microsoft Foundry, Databricks and several automation platforms are examples given. Using tools you can access, practice packaging a small workflow and documenting deployment, monitoring and rollback steps rather than treating every named product as mandatory.

• Build and test an AI workflow

  As preferred preparation, use Python to build a small AI/ML or LLM workflow, add repeatable tests and record a quality measure. Python and LLM integrations are valued in the posting, but this exercise is a preparation suggestion, not a specified project.

• Exercise agent and retrieval controls

  Agent tool use, API orchestration, retrieval pipelines and AI security controls are preferred experience. As a practice exercise, sketch or prototype a bounded retrieval-and-tool workflow, then test an unauthorized request, a poor retrieval result and a prompt-injection attempt; document access decisions and audit events.

• Lead a cross-functional delivery discussion

  The role partners with business, technology, architecture and security teams to identify opportunities and deliver solutions. Rehearse a short discovery conversation for a hypothetical automation request: elicit requirements and data needs, surface risks, propose an iteration plan and state how value would be measured.

• Map your delivery evidence to the requirements

  Prepare a one-page evidence map for the stated requirements: at least eight years of progressive AI delivery, productionized solutions, impact across multiple functions, and measurable business value. Use only experience you can substantiate.

• Build an end-to-end project walkthrough

  Select a real AI or automation project and outline the business problem, your hands-on and leadership contributions, systems integrated, production architecture, results, and trade-offs. Highlight enterprise-platform integration where your experience supports it.

• Practice a production architecture exercise

  Sketch a hypothetical AI solution for one function named in the posting, such as supply chain or finance. Show integration boundaries, scalability and security choices, deployment, and MLOps or LLMOps controls; label assumptions rather than presenting the design as a Solstice project.

• Prepare a reliability and risk review

  Draft a short review covering evaluation metrics, observability, failure handling, data classification, access controls, and a responsible-AI approval trade-off. The role calls for rapid iteration while helping the Corporate AI Risk Council balance risk, compliance, and speed.

• Inventory relevant tools and technical depth

  Create a concise matrix of AI, automation, and enterprise platforms you have actually used, your level of involvement, and a concrete outcome for each. Prioritize the posting’s required multi-tool proficiency; distinguish preferred experience such as agents, retrieval pipelines, and Python.

• Rehearse a cross-functional delivery example

  Prepare an example of working with business, architecture, technology, and security stakeholders to define requirements and deliver a solution. Include how you explained technical choices and resolved competing priorities; add manufacturing or managed-service-provider experience only if applicable.

• Write targeted questions for the employer

  Bring questions about priority use cases, existing AI and automation platforms, production support ownership, evaluation standards, and how the AI Risk Council participates in decisions. These are topics to clarify, not assumptions about Solstice’s current processes.