Agentic AI Engineer

Verizon is looking for an Agentic AI Engineer to design and deliver autonomous, multi-agent workflows that go beyond static text generation. In this hybrid role based in Irving, TX, you will connect large language models to enterprise data sources and real-world APIs while keeping systems safe, observable, and governed.

The work focuses on building stateful agent behavior, long-running orchestration, and reliable integrations across databases and external services, with an emphasis on performance monitoring and preventing unwanted model behavior such as hallucinations, infinite loops, and drift.

What you’ll do

• Design stateful, multi-step agentic workflows using reasoning frameworks such as ReAct, Plan-and-Solve, and Reflection.
• Build execution layers that connect LLMs to external systems using REST APIs, database connectors, and standards including Model Context Protocol (MCP).
• Implement short-term and long-term memory architectures, including state persistency, context window optimization, and hybrid RAG retrieval layers.
• Create offline test sets (“golden sets”) and real-time observability pipelines using LLM-as-a-judge, trace analysis, and latency/cost monitoring to reduce hallucinations, infinite loops, and drift.
• Containerize and deploy multi-agent microservices with scalable async patterns, distributed queues, and fallback mechanisms.

What you’ll need

• A Bachelor’s degree or 4+ years of work experience (including work and/or military experience or specialized training).
• 4+ years of relevant experience.
• Experience building Multi-Agent Systems with complex delegation models, including human-in-the-loop escalation paths.
• Hands-on experience developing enterprise-grade agentic workflows, multi-agent topologies, and orchestration pipelines using platforms such as LangGraph, Google ADK / CES, Sierra, Decagon, or similar tools.
• Experience with LLMs, prompt engineering, RAG architectures, and generative AI guardrails.

Technologies you may work with

• ReAct, Plan-and-Solve, Reflection, Model Context Protocol (MCP), LLM-as-a-judge
• LangGraph, Google ADK / CES, Sierra, Decagon
• Large Language Models (LLMs), prompt engineering, retrieval-augmented generation (RAG)
• Containerization, REST APIs, database connectors, distributed queues
• Vertex AI Search, Enterprise Search, Salesforce, Genesys
• Python, Java, NodeJS
• REST/SOAP APIs, pub/sub messaging, webhook integrations
• Google AI, telephony adapters

Where you’ll work and schedule

• This hybrid position has a defined work location with work from home and a minimum of three days per week in the office, set by your manager.
• Weekly hours: 40

Compensation and benefits

• Salary range: USD 101,000 - 194,000 per year for the listed location(s), based on a full-time schedule.
• The position is incentive based with the potential to earn more.
• Medical, dental, vision
• Short and long term disability
• Basic life insurance; supplemental life insurance; AD&D insurance
• Identity theft protection
• Pet insurance
• Group home & auto insurance
• Matched 401(k) savings plan
• Up to 8 company paid holidays per year; up to 6 personal days per year
• Paid parental leave
• Adoption assistance
• Tuition assistance
• Premium pay such as overtime, shift differential, holiday pay, allowances (depending on the role)