• Build and maintain AI-powered data pipelines and extraction processes (batch and streaming) from internal relational data sources and unstructured documents (PDF, Word, PowerPoint) into structured datasets within Databricks. • Design and manage text embeddings and vector stores within Databricks for use with vector indexing and retrieval solutions. • Design, develop, and maintain custom tools implemented as MCP servers and Databricks applications to extend agent and model capabilities. • Design, develop and implement AI Agents using frameworks like LangChain and LangGraph. • Implement LLM scorers to validate and monitor agents, applications and models. • Prevent issues like hallucinations or unnecessary actions through structured testing and guardrails. • Drive continuous improvement through prompt engineering, pipeline optimization, vector store tuning, and scorer refinement to ensure high-quality LLM responses. • Collaborate on production deployments, monitoring, and scalability of ML and LLM-based services.

The Artificial Intelligence (AI) Solutions Architect (Contractor) will work with our existing internal AI product management team, Vendor Teams and existing development staff to develop, enhance and streamline our AI systems. Lead the design and implementation of secure, scalable AI and Agentic AI solutions aligned with DES mission, policies, and public-sector standards. Define end-to-end AI architectures including data ingestion, RAG pipelines, LLM orchestration, and human-in-the-loop workflows. Provide technical leadership, relevant architectural documentation for approvals, and guidance to internal teams and system integrators. Ensure compliance with State of Arizona security, privacy, accessibility, and data-governance requirements (PII, HIPAA, CJIS where applicable). Establish reference architectures and best practices for responsible, ethical, and compliant AI use across DES. Evaluate and integrate AI platforms, cloud services, and vendor solutions in coordination with procurement and legal teams. Design architectures that support explainability, auditability, bias mitigation, and transparency in decision-support systems. Collaborate with DevOps (cloud ops), cybersecurity, and data teams to operationalize AI solutions from proof of concept to production systems. Support AI strategy, roadmap development, and executive briefings to scale AI capabilities across DES programs.

• Stand up and run squads (Discovery → Prototype → Product → Platform & SRE). • Design and ship RAG/agent systems: pick models (e.g., Anthropic Claude, OpenAI, Google, or open-weights like Llama/Mistral), define tools/functions, and choose retrieval (default Postgres + pgvector, scale to Weaviate/Qdrant/Pinecone when needed). • Operate AI safely: evals & guardrails, structured outputs (JSON/Schema), PII redaction, refusal policies, cost/latency budgets, and LLM observability. • Own delivery outcomes: SLOs, quality, cost, velocity; release with feature flags and canaries. • Be client-facing: discovery, scoping, SoW, roadmap, QBRs. • Hire/coach Tech Leads, EMs, and PMs; level up practices.

• Design and development of an agent-based AI system • Implementation of LLM-based workflows with LangChain and LangGraph • Building and optimizing RAG pipelines • Integration of Azure AI Services into existing cloud infrastructures • Infrastructure automation with Terraform • Close collaboration with business units and cloud architecture teams • Ensuring scalability, stability, and security of the solution