AI Systems Engineer - Scalable Agent Orchestration

Socket.dev; New York (NY); On-site; USD 158,000 - 214,000; Full time

Benefits offered by this job

RSUs

401(k) matching

Health insurance

Paid time off

Parental leave

Job summary

Amazon Ads in New York seeks an experienced software engineer to help build the agent execution engine for autonomous campaigns. You’ll design core infrastructure for LLM reasoning, multi-step orchestration, and RL that enables agents to plan, reason, and act across advertiser workflows.

You will collaborate with senior engineers and scientists to productionize research prototypes, focusing on tool integration, distributed systems, and scalable cloud-native solutions at massive scale.

Qualifications

• 3+ years of non-internship software development experience.
• 2+ years of design or architecture experience.
• Experience with at least one programming language and AI/ML technologies.

Responsibilities

• Implement core services and APIs for agent orchestration, tool invocation, and state management.
• Build scalable data and tool registries enabling cross-workflow reasoning.
• Create evaluation pipelines (offline/online) to measure reasoning quality and outcomes.
• Develop infrastructure for fine-tuning and reinforcement learning pipelines with scientists.
• Ensure performance, scalability, observability, and maintainability through strong practices.
• Participate in system design, code reviews, and team architecture discussions.
• Continuously learn and apply new techniques in agentic frameworks and cloud-native dev.

Skills

Software development

System design

Programming languages

AI/ML experience

Education

Bachelor's degree in CS

Job description

Amazon Ads in New York seeks an experienced software engineer to help build the agent execution engine for autonomous campaigns. You’ll design core infrastructure for LLM reasoning, multi-step orchestration, and RL that enables agents to plan, reason, and act across advertiser workflows.

You will collaborate with senior engineers and scientists to productionize research prototypes, focusing on tool integration, distributed systems, and scalable cloud-native solutions at massive scale.