Applied AI Engineer, Agent Harness

Build assistant-quality behavior with a privacy-forward, eval-driven agent harness in Apple’s Siri organization. This onsite role in Cupertino focuses on turning large language models into assistants that can take action, balancing low latency with reliability across on-device and server systems.

You’ll work on the agent runtime that powers planning, session lifecycle, tool execution, and interruption mid-response, with continuous validation using latency metrics and evaluation workflows.

What you’ll do

• Design and iterate core components of the agent runtime, including the planning loop, session lifecycle, concurrency, and cancellation
• Build the transcript and conversation state, plus the platform layer that defines tools and dynamic instructions delivered to the model
• Develop a reliable tool executor that dispatches tool calls into Apple apps and services, handling results and failures
• Integrate on-device and server models, implementing low-latency streaming and supporting interruption mid-response
• Embed safety, security, and privacy protections into harness components in collaboration with security and privacy partners
• Validate harness updates using evals and latency metrics, adapting the harness as new model generations arrive
• Create tracing, replay, and testing tools to make a nondeterministic system debuggable

Key requirements

• Bachelor’s degree in Computer Science or a related field, plus 2+ years of industry experience
• Strong programming ability in a systems language such as Swift, C++, Rust, Objective-C, or Go
• Solid foundation in concurrency, inter-process communication, and performance optimization
• Hands-on experience building LLM-powered systems including agents, tool calling, or prompt and context design (professional work or substantial personal projects)
• Daily use of agentic coding tools such as Claude Code, Codex, or Pi, with informed opinions on what helps harnesses succeed or fail
• Ability to communicate clearly and make progress on ambiguous problems

Preferred qualifications

• Experience with Swift and Apple platform development
• Experience building agent harnesses, orchestration frameworks, or multi-step tool-use systems
• Experience integrating model inference, streaming protocols, or latency-sensitive client-server systems
• Experience using evals to diagnose and improve LLM behavior
• Experience designing APIs or platforms adopted by other engineering teams
• Familiarity with privacy and security engineering in production systems
• Active engagement with current research and practice in agent harness design, including context engineering, tool use, and long-horizon agents, with a track record of turning ideas into practical improvements

Compensation and location

Cupertino, CA (onsite)
Salary range: USD 150,400 - 225,300 per year.

Technologies you may work with

Swift, C++, Rust, Objective-C, Go, plus Claude Code, Codex, and Pi.

Benefits

• Comprehensive medical and dental coverage
• Retirement benefits
• Discounted products and free services
• Reimbursement for certain educational expenses, including tuition
• Opportunity to become an Apple shareholder through discretionary employee stock programs
• Eligible for discretionary restricted stock unit awards
• Ability to purchase Apple stock at a discount through the Employee Stock Purchase Plan (voluntary participation)
• May be eligible for discretionary bonuses or commission payments and relocation (as applicable)