Principal Architect: Autonomous AI Safety & Controls

Verizon; Basking Ridge (NJ); Hybrid; USD 121,000 - 231,000; Full time

Job summary

Verizon is seeking a Principal hands-on architectural engineer and technical lead to build and optimize enterprise AI control planes. You will connect AI developer platforms with centralized governance hubs, focusing on scalable, secure, multi-agent execution at Fortune 20 scale.

You will design end-to-end pipelines, model inference optimization, and CI/CD guardrails while collaborating with platform engineers and product teams in a hybrid work setting.

Qualifications

• Bachelor’s degree or 4+ years of work experience.
• 6+ years of relevant AI/architectural experience.
• 4+ years designing multi-agent frameworks and MCP/A2A communications.
• Deep experience in LLM operationalization, evaluation pipelines, and automated safety testing.
• Experience lowering LLM inference costs and throughput with classifiers or routing.
• Experience in cloud architecture (AWS/GCP) and CI/CD automation.
• Experience defending agentic ecosystems against prompt injection and data exfiltration.

Responsibilities

• Architect agentic evaluation pipelines for continuous validation and observability.
• Lead fine-tuning of LLMs to improve throughput and reduce costs.
• Implement RL techniques to optimize autonomous agent accuracy.
• Design semantic routing and classification to minimize frontier LLM calls.
• Automate governance with CI/CD pipelines integrating guardrails.
• Collaborate with AI Platform Engineers and DevOps/DataOps to integrate models.
• Design IAM and boundary logic within the Agent Gateway for secure workflows.

Skills

AI Engineer

Architect

LLM Inference

CI/CD

Security

IAM

KAIROS

AEGIS

AWS Bedrock

GCP

Education

Bachelor's degree

Tools

Atlas AI

KAIROS

AEGIS

OpenTelemetry

Langfuse

Arize

JEV

Job description

Verizon is seeking a Principal hands-on architectural engineer and technical lead to build and optimize enterprise AI control planes. You will connect AI developer platforms with centralized governance hubs, focusing on scalable, secure, multi-agent execution at Fortune 20 scale.

You will design end-to-end pipelines, model inference optimization, and CI/CD guardrails while collaborating with platform engineers and product teams in a hybrid work setting.