Agentic AI Engineer

Integra Mission Critical•Dallas, TX

1d•Onsite

About The Position

The Agentic AI Engineer builds production-ready, AI-powered solutions that address operational challenges across Manufacturing, Construction, Finance, Project Controls, and Operations. The role focuses on agents that evaluate information, identify conditions and exceptions, make decisions, and initiate workflows or other actions across enterprise systems—not merely answer questions. The engineer partners with business teams and technology stakeholders to identify automation opportunities and deliver secure, governed, measurable solutions. The position is 100% onsite in the North Dallas/Plano area. The engineer designs and maintains single-agent and multi-agent solutions using Microsoft Azure AI, Copilot technologies, and code-first orchestration frameworks. Work includes implementing agent planning, tool use, memory, context management, routing, delegation, human review, bounded autonomy, approvals, escalation, rollback, and other controls. The role also develops reusable components, connectors, policies, templates, and reference architectures. The position includes building retrieval-augmented generation solutions grounded in trusted enterprise data; integrating agents with business applications and enterprise platforms; and implementing secure, auditable actions through APIs, workflows, and approved automation services. The engineer establishes evaluation, responsible AI, security, governance, monitoring, and operational practices, and collaborates across business, engineering, security, compliance, and data-governance teams. Successful solutions are piloted, evaluated, and transitioned into governed production services, with performance and business value monitored over time.

Requirements

Bachelor's degree in Computer Science, Software Engineering, Data Science, Artificial Intelligence, Information Systems, or a related discipline, or equivalent professional experience.
5+ years of experience in software engineering, AI engineering, data engineering, automation, or enterprise application development.
2+ years of hands-on experience building generative AI, retrieval-augmented generation, copilots, or agent-based solutions.
Proven experience moving AI prototypes into secure, monitored, production-grade enterprise solutions.
Strong Python or C# development skills and experience integrating AI solutions with APIs, enterprise data, workflows, and business applications.
Strong analytical, process-discovery, facilitation, documentation, and stakeholder-management capabilities.
Knowledge of Microsoft Foundry, Foundry Agent Service, Azure OpenAI, Microsoft Copilot Studio, Azure AI Search, and Azure AI Content Safety.
Experience with OpenAI GPT and Anthropic Claude models and agent orchestration using Microsoft Agent Framework, Semantic Kernel, or comparable frameworks.
Knowledge of prompt engineering, structured outputs, function calling, tool use, context management, memory, multi-agent patterns, evaluation, guardrails, safety testing, and production monitoring.
Knowledge of retrieval-augmented generation, embeddings, vector search, semantic and hybrid retrieval, reranking, document ingestion, indexing, permissions filtering, and citations.
Experience with Python and/or C#/.NET, REST APIs, OpenAPI, JSON, asynchronous programming, SQL, NoSQL concepts, caching, Docker, serverless architectures, Git, CI/CD, and automated testing.
Knowledge of Microsoft Entra ID, managed identities, Azure Key Vault, RBAC, Azure Monitor, Application Insights, Log Analytics, Microsoft Fabric, and Power BI.

Nice To Haves

Experience integrating AI with ERP, project controls, construction management, manufacturing, CRM, HCM, procurement, content, and analytics platforms.
Practical knowledge of manufacturing, supply chain, construction, project controls, finance, ERP, workflow analysis, and continuous improvement.
Experience with Oracle Fusion ERP, Deltek Costpoint, SAP S/4HANA, Microsoft Dynamics 365, Primavera P6, Procore, Autodesk Construction Cloud, Salesforce, ServiceNow, SharePoint, Teams, Azure Databricks, or ADLS Gen2.
Ability to distinguish work that should be assisted, automated, escalated, or retained under human control.

Responsibilities

Design, build, deploy, and maintain enterprise AI agents, copilots, conversational applications, and intelligent workflow solutions.
Create single-agent and multi-agent architectures using managed agent services, code-first frameworks, and Microsoft Copilot technologies.
Implement agent planning, tool use, function calling, memory, context management, routing, delegation, reflection, and human-in-the-loop controls.
Design bounded autonomy, approval checkpoints, escalation paths, timeout behavior, rollback, and compensation mechanisms.
Create reusable agent components, prompt libraries, tools, connectors, policies, templates, and reference architectures.
Build solutions using Microsoft Foundry, Foundry Agent Service, Azure OpenAI, Azure AI Search, Azure AI Content Safety, and related Azure AI capabilities.
Develop enterprise agents using Microsoft Copilot Studio and integrate them with Microsoft 365, Teams, Power Platform, and approved enterprise channels.
Use Microsoft Agent Framework, Semantic Kernel, or comparable frameworks for code-first agent development.
Evaluate and route approved models based on quality, latency, cost, safety, reliability, data handling, and task fit; implement model abstraction, fallback, retry, timeout, graceful degradation, structured output, function calling, and governed model-version management.
Design retrieval-augmented generation solutions grounded in trusted enterprise data and documents.
Build indexing, chunking, metadata, embedding, hybrid search, semantic search, reranking, citation, and access-filtering pipelines.
Integrate structured and unstructured knowledge from Microsoft Fabric, data lakes, databases, SharePoint, enterprise repositories, APIs, and SaaS systems.
Ensure AI responses are grounded, traceable, current, permission-aware, and supported by relevant source references.
Work with data owners and stewards to define authoritative sources, business terminology, metadata, and knowledge-quality standards.
Partner with business leaders and subject-matter experts to identify repetitive work, decision bottlenecks, inefficiencies, control gaps, and automation opportunities.
Map current- and future-state processes, including decisions, exceptions, roles, data inputs, systems, business rules, and expected outcomes.
Prioritize use cases based on business value, feasibility, risk, data readiness, adoption, operating cost, and measurable return.
Define when human review is required before an agent communicates externally, changes a system of record, commits funds, or affects operations.
Create pilots, validate outcomes, measure benefits, and transition successful solutions into governed production services.
Build secure tools and connectors for agents to interact with enterprise applications, APIs, workflows, databases, message platforms, and approved automation services.
Integrate agents with applicable enterprise platforms, including ERP, project controls, construction management, manufacturing, CRM, Microsoft Fabric, and Power BI systems.
Use Azure API Management, Azure Functions, Logic Apps, Service Bus, Event Grid, and containerized services to expose governed agent actions.
Apply least privilege, user-delegated authorization, transaction boundaries, idempotency, validation, rate limits, audit logging, and business-rule enforcement.
Ensure agents do not bypass system-of-record controls, approval processes, data entitlements, or enterprise integration standards.
Define evaluation frameworks for groundedness, relevance, correctness, completeness, tool selection, task completion, consistency, safety, latency, and cost.
Create representative evaluation datasets and automated regression tests for prompts, models, retrieval, tools, workflows, and agent behavior.
Conduct red-team testing for prompt injection, data exfiltration, unauthorized actions, harmful outputs, hallucinations, and tool misuse.
Apply responsible AI, privacy, security, compliance, transparency, accountability, and human-oversight principles throughout the solution lifecycle.
Implement content filtering, access controls, data-loss prevention, audit trails, release gates, canary rollouts, and rollback capabilities.
Implement telemetry for prompts, responses, retrieved context, tool calls, workflow steps, approvals, errors, token usage, latency, cost, and user feedback.
Use Azure Monitor, Application Insights, Log Analytics, and AI-specific tracing and evaluation capabilities.
Monitor quality drift, retrieval degradation, model changes, safety events, tool failures, adoption, and realized business value.
Troubleshoot production AI solutions, perform root-cause analysis, and implement permanent corrective actions.
Create dashboards, alerts, runbooks, support procedures, service-level expectations, incident processes, and recovery plans.
Collaborate with business and technology teams to define AI use cases, acceptance criteria, delivery roadmaps, and measurable value.
Partner with Data Engineering, Integration, DevSecOps, Cybersecurity, Legal, Compliance, Risk, Internal Audit, and data-governance stakeholders.
Communicate AI capabilities, limitations, risks, and trade-offs to technical and nontechnical stakeholders.
Create documentation, demonstrations, training materials, patterns, and standards to improve enterprise AI literacy.
Mentor engineers and contribute to an enterprise community of practice for applied and agentic AI.