Senior AI Agentic Engineer
Truist Financial
- Location
- Charlotte, NC
- Work model
- On-Site
- Level
- Senior
- Posted
- 1d ago
Skills
About this role
The position is described below. If you want to apply, click the Apply Now button at the top or bottom of this page. After you click Apply Now and complete your application, you'll be invited to create a profile, which will let you see your application status and any communications. If you already have a profile with us, you can log in to check status. Need Help? If you have a disability and need assistance with the application, you can request a reasonable accommodation. Send an email to Accessibility (accommodation requests only; other inquiries won't receive a response). Regular or Temporary: Regular Language Fluency: English (Required) Work Shift: 1st shift (United States of America) Please review the following job description: The Senior AI Agentic Engineer is a senior production builder responsible for designing, implementing, testing, and operating enterprise-grade agentic systems that solve real business problems at scale. This role sits at the center of the Forge delivery model, translating approved architecture patterns into production workflows that combine prompts, tools, APIs, memory, retrieval, orchestration logic, and human oversight into working enterprise capabilities. This is a hands-on senior engineering role focused on agent workflows, multi-step reasoning, prompt architecture, tool integration, memory/context management, retrieval-augmented generation (RAG), evaluation, observability, and governed deployment. The engineer works across platforms and frameworks used to deliver production AI capabilities, including Microsoft and Azure-based agent platforms, enterprise workflow tooling, and other approved agentic engineering stacks where needed. Daily work includes building agent flows, implementing orchestration logic, integrating APIs and enterprise systems, tuning prompts and execution patterns, validating output quality, hardening reliability, instrumenting observability, and supporting secure deployment through enterprise release controls. The role is expected to operate with senior-level ownership, production judgment, and deep technical accountability while still moving at startup speed inside enterprise governance. Essential Duties and Responsibilities Following is a summary of the essential functions for this job. Other duties may be performed, both major and minor, which are not mentioned below. Specific activities may change from time to time. Design, build, and deploy production-ready AI agents and agentic workflows that combine prompts, tools, APIs, memory, retrieval, and business logic into useful enterprise capabilities. Implement single-agent and multi-agent workflows using approved Forge patterns, including tool orchestration, supervisor-worker patterns, human-in-the-loop controls, session-aware execution, and safe fallback behaviors. Translate architecture and product intent into robust implementations that are observable, testable, secure, and maintainable in enterprise production environments. Build and maintain integrations to enterprise systems, APIs, workflow services, and governed data sources that enable agents to act safely and effectively within approved boundaries. Design and maintain prompt architectures, tool schemas, retrieval logic, memory behavior, and structured execution controls that improve agent accuracy, reliability, and explainability. Contribute to RAG, grounding, and knowledge-retrieval patterns using search, vector stores, document pipelines, and governed enterprise data services. Implement runtime controls for resilience and safety, including error handling, retries, tool constraints, structured outputs, human escalation, and observable execution behavior. Support evaluation and regression validation for agentic solutions, including prompt changes, model changes, workflow changes, and quality benchmarks tied to business outcomes. Use approved engineering acceleration tools to speed code generation, testing, refactoring, and