AI Agent Control Architect - Senior Manager

EY; McLean (VA); Hybrid; USD 126,000 - 230,000; Full time

Benefits offered by this job

Hybrid work model

Comprehensive benefits package

Generous paid time off

Job summary

EY is seeking an experienced agent AI Software Architect to design and implement AI Agent Control, including tool usage, state management, execution environment control and multiagent APIs. You will lead synthesis models, authentication architectures and benchmarking against evals while collaborating with AI, Security, Data and Infrastructure teams.

The role requires advancing post-training environments, safety, observability and IP development, with a flexible hybrid work model and EY’s global

Qualifications

• Advanced degree in Electrical Engineering, Computer Science or Mathematics.
• 5+ years of experience in software engineering and AI.
• Experience working with multi-agent system administration and modern secrets management.
• Use of AI native software development tools / code assist (Cursor, Claude Code, Qodo/Codium, Gemini, AWS Q, GitHub)
• Familiarity with scripting languages (Python, Bash, Typescript etc….) and software languages (Go, Rust, Java etc…).
• Understand AI Native design process and principles
• Experience with cryptographic methods of identity and auth.
• Experience with modern software IDE, CI/CD pipeline, modern code hosting and collaboration (Github/Gitlab/Azure DevOps etc…)
• Experience with Nvidia NVAIE, AWS and MSFT Azure AI stacks.

Responsibilities

• Design architectures for agent identity and authentication in large scale multi-agent enterprise systems.
• Design of post training and tuning/customization environments.
• Design of synthesis models for AI open source models.
• Implement and test these architectures benchmarking performance against agent and human designed evals.
• Design comprehensive AI Agent control frameworks.
• Drive agent design requirements and implement through AI Native design process.
• Assess through primary and secondary research emerging agent authorization architectures and frontier model proposals.
• Collaborate with other architectural experts in AI, Security, Data and Infrastructure to insure balanced and practical outcomes.
• Support strategic business development activities for emerging technology as required.
• Present at academic conferences and support development of IP filings
• Hands on development of the advanced AI capabilities in agentic control.

Skills

Analytical skills

Leadership

Cross-functional leadership

Partnership building

Emotional intelligence

Critical thinking

Stakeholder alignment

Education

Advanced degree in Electrical Engineering, Computer Science or Mathematics

Tools

Cursor

Claude Code

Qodo/Codium

Gemini

AWS Q

GitHub

Job description

EY is seeking an experienced agent AI Software Architect to design and implement AI Agent Control, including tool usage, state management, execution environment control and multiagent APIs. You will lead synthesis models, authentication architectures and benchmarking against evals while collaborating with AI, Security, Data and Infrastructure teams.

The role requires advancing post-training environments, safety, observability and IP development, with a flexible hybrid work model and EY’s global