Lead AI Agent Control Architect

EY; Phoenix (AZ); Hybrid; USD 150,000 - 230,000; Full time

Benefits offered by this job

Hybrid work model

Flexible vacation policy

Total Rewards package

Job summary

EY is seeking an experienced Lead AI Software Architect to lead a team of domain experts in designing AI Agent Control. You will shape architectures for multi-agent systems, tool usage, state management, and policy application in a fast-paced environment.

You will drive integration across AI, Security, Data and Infrastructure, mentor engineers, and collaborate with academia to advance agent authentication, observability, and federated learning.

Qualifications

• Advanced degree in a relevant engineering/science field and 10+ years of software/AI experience.
• Experience with multi-agent systems, agent authentication, and orchestration.
• Proficiency with AI-native tools, modern CI/CD and secure software practices.

Responsibilities

• Lead AI software architecture for universal agentic control harnesses.
• Design and implement multi-agent control frameworks and authentication.
• Mentor engineers and collaborate with AI, security and data teams.
• Drive post-training and tuning environments and governance.
• Present research at conferences and support IP development.

Skills

Leadership

Architectural design

AI/ML expertise

Mentoring

Cross-functional collaboration

Multi-agent systems

Critical thinking

Education

Advanced degree in Electrical Engineering, Computer Science or Mathematics

Tools

Cursor

Claude Code

Qodo/Codium

Gemini

AWS Q

GitHub

Python

Bash

TypeScript

Go

Rust

Java

NVIDIA NVAIE

AWS AI Stack

Azure AI Stack

Job description

EY is seeking an experienced Lead AI Software Architect to lead a team of domain experts in designing AI Agent Control. You will shape architectures for multi-agent systems, tool usage, state management, and policy application in a fast-paced environment.

You will drive integration across AI, Security, Data and Infrastructure, mentor engineers, and collaborate with academia to advance agent authentication, observability, and federated learning.