AI Agent Infra Engineer: LLM Pipelines & Deployment

Socket.dev; Cupertino (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Apple is seeking talented engineers to build the infrastructure and agentic systems for prototype, optimize, and ship Apple Intelligence features. You will join a collaborative team focused on LLMs, agent harnesses, evaluation, and pipelines connecting training, adaptation, and deployment.

Successful candidates bring a strong software and systems background, hands-on experience building ML infrastructure end to end, and the judgment to trade velocity for maintainability.

Qualifications

• Proven ability to scope and lead complex, multi-month projects from concept to production.
• Define goals and deliver results amid uncertainty in AI product development.
• Experience designing, building, and operating infrastructure for ML—training/evaluation pipelines, data/experiment tooling, serving, or agent harnesses.
• Strong software engineering fundamentals: distributed systems, APIs, testing, maintainable code.
• Strong Python and UNIX skills and experience with agentic coding tools.

Responsibilities

• Build infrastructure and agent systems used to prototype, optimize, and ship Apple Intelligence features.
• Collaborate with software developers and applied scientists on LLMs, agent harnesses, and evaluation pipelines.
• Navigate pragmatic tradeoffs between velocity and long-term maintainability in AI product development.

Skills

Python

UNIX

Distributed systems

APIs

Testing

Maintainable code

Leadership

Education

BS + 5+ years (or MS + 3+ years, or PhD + 1+ year)

Tools

Agent frameworks

LLM evaluation tooling

ML training/evaluation pipelines

Job description

Apple is seeking talented engineers to build the infrastructure and agentic systems for prototype, optimize, and ship Apple Intelligence features. You will join a collaborative team focused on LLMs, agent harnesses, evaluation, and pipelines connecting training, adaptation, and deployment.

Successful candidates bring a strong software and systems background, hands-on experience building ML infrastructure end to end, and the judgment to trade velocity for maintainability.