Senior AI Architect - Generative AI

The Senior AI Architect, Generative AI is the technical authority responsible for defining how foundation-model, RAG, conversational, and agentic solutions are designed, evaluated, integrated, governed, and operated across Wind Engineering.

This role creates architectures and reusable patterns for enterprise knowledge systems, engineering copilots, document intelligence, automated workflows, and multi-agent applications while ensuring grounding, traceability, security, human oversight, and measurable value.

The architect separates compelling demonstrations from dependable engineering systems by establishing disciplined model selection, retrieval design, evaluation, guardrails, observability, and ownership practices.

• Define reference architectures for foundation-model applications, RAG, tool-using assistants, agentic workflows, document intelligence, and multimodal engineering use cases.
• Establish approved patterns for model gateways, prompt and context management, retrieval, memory, tool invocation, orchestration, evaluation, logging, and user experience.
• Define selection criteria for hosted, enterprise, open-weight, fine-tuned, and task-specific models based on quality, security, latency, cost, controllability, and supportability.
• Ensure solutions align with approved GE Vernova platforms, identity, data, integration, and deployment standards.

• Define when deterministic automation, a single assistant, or an agentic architecture is appropriate.
• Establish patterns for planning, tool use, state, memory, handoffs, retries, approvals, termination, and exception handling.
• Integrate agents with engineering applications, APIs, workflow engines, data systems, simulation tools, and enterprise services using least-privilege access.
• Prevent uncontrolled autonomy by defining bounded actions, approval gates, transaction limits, and recoverable failure behavior.

• Define evaluation frameworks covering task success, groundedness, retrieval quality, factuality, completeness, instruction adherence, safety, latency, cost, and user value.
• Create representative test sets from actual engineering tasks, including difficult, ambiguous, adversarial, and out-of-scope cases.
• Establish automated, model-based, deterministic, and expert human evaluation where each is appropriate, with calibrated use of LLM-as-judge methods.
• Require release gates, regression testing, red-team evidence, and post-deployment monitoring before broad adoption.

• Design GenAI experiences that fit engineering decision points, standard work, document flows, and existing applications rather than creating disconnected chat interfaces.
• Define patterns for human review, structured output, source evidence, approvals, write-back, and exception handling.
• Ensure generated content entering engineering records is identifiable, reviewable, and subject to appropriate validation.
• Partner with process owners to measure cycle-time, quality, adoption, and rework impacts.

• Mentor Embedded AI Architects, AI engineers, data scientists, and product teams in GenAI system design and evaluation.
• Lead architecture reviews, evaluation clinics, red-team exercises, technical demonstrations, and reusable-pattern workshops.
• Define competency expectations across RAG, agents, evaluation, security, prompt/context engineering, and workflow integration.
• Maintain a pragmatic view of emerging models and frameworks, adopting capabilities based on measured engineering value.

• Bachelor's degree in Computer Science, Engineering, Applied Mathematics, Data Science, or a related technical field; advanced degree strongly preferred.
• Significant hands-on experience architecting and deploying production generative AI, RAG, conversational, or agentic systems.
• Demonstrated knowledge of foundation models, embeddings, retrieval, reranking, prompt/context engineering, structured generation, tool use, and model evaluation.
• Experience with orchestration and application frameworks such as LangGraph, Semantic Kernel, n8n, or equivalent, without dependence on any single framework.
• Experience integrating AI applications with enterprise APIs, identity, data platforms, document repositories, workflow systems, and observability services.

• Experience designing test sets and evaluation pipelines for quality, groundedness, safety, regressions, latency, and cost.
• Understanding of GenAI security risks, including prompt injection, data leakage, insecure tool use, and untrusted content.
• Ability to define enterprise standards, document decisions, mentor teams, and communicate architecture trade-offs to technical and non-technical stakeholders.
• Experience with engineering, scientific, manufacturing, service, or regulated knowledge workflows.
• Experience with Azure OpenAI, Amazon Bedrock, Google Vertex AI, or comparable enterprise model platforms and model gateways.
• Experience with vector and hybrid search, knowledge graphs, document parsing, multimodal models, fine-tuning, adapters, or synthetic-data generation.
• Experience implementing agent evaluation, runtime tracing, policy enforcement, and approval-based action execution.
• Familiarity with Responsible AI frameworks, model risk management, cybersecurity threat modeling, and audit evidence for AI systems.
• Experience evaluating commercial GenAI platforms and balancing delivery speed against control, portability, ownership, and total cost.