Staff AI Architect, Family AI Lab

This is a fully remote position, open to applicants in United States.

📋 Description

• Design the inference pipeline that transforms Life360 into a context layer for LLMs to reason over: soul files, family context, behavioral signals, and document understanding.

• Select models and hosting options. Determine what should be self-hosted, API-based, fine-tuned, or prompt-engineered, and reassess these choices as the landscape evolves.

• Develop the serving infrastructure, including latency budgets, batching, caching, fallbacks, and graceful degradation. Ensure it operates at scale before the need arises.

• Construct the evaluation loop. Assess whether changes enhance the product or detract from it, rather than just confirming the demo's functionality. This distinguishes a genuine product from a mere magic trick.

• Manage the cost model. Monitor expenditures per user, feature, and family. Identify the levers (context compression, model routing, pre-computation) that maintain viable unit economics from a test group to millions.

• Determine what data should be preserved, summarized, or discarded. The architectural decision regarding "what our data appears like to an LLM" rests with you.

• Establish standards for safety, privacy, and trust at the infrastructure level. Sensitive categories must be filtered before dissemination. Ensure encryption, access controls, and audit trails are integrated, not added later.

⛳️ Requirements

• Extensive engineering experience, including significant time spent building and managing AI systems in production at substantial scale. Experience with real users, real loads, and being on-call.

• Proficient in modern LLM serving (vLLM, TGI, SGLang, hosted APIs — whatever is appropriate for the task). No rigid adherence to a single method.

• You think in evaluations. You've created harnesses and understand the distinction between vibes-based iteration and tangible progress, having experienced the frustration of lacking one.

• Capable of building a cloud system from scratch. Select services, integrate them, ensure reliability, and manage costs effectively.

• Have strong opinions about cost management. You've witnessed inference bills spiral out of control and have instincts for identifying sources of inefficiency.

• AI-native in your daily tasks. Experienced with Claude Code, Cursor, or similar tools. You intuitively navigate agentic workflows, prompt engineering, context window management, and MCP/function calling.

• Comfortable making decisions with incomplete information. The specifications will evolve, and you don't require them to be finalized before beginning construction.

• Excellent written communication skills. You produce specs, decision records, and playbooks that can be acted upon precisely by both agents and humans.

• Familiarity with agent systems, retrieval pipelines, or long-context personalization.

• Experience in privacy architecture within multi-user systems, where context is shared among individuals with varying access rights.

• You have developed a project where the data itself served as a competitive advantage, and you recognize how this differs from building on publicly available data.

🏝️ Benefits

• Comprehensive medical, dental, vision, life, and disability insurance plans (100% covered for US employees). Supplemental medical and dental plans are available for Canadian employees.

• 401(k) plan with a company matching program for US employees and an RRSP with a DPSP plan for Canadian employees.

• Employee Assistance Program (EAP) dedicated to mental wellness.

• Flexible PTO and 12 company-wide days off throughout the year.

• Opportunities for Learning & Development programs.

• Provision of equipment, tools, and reimbursement support to foster a productive remote work environment.

• Complimentary Life360 Platinum Membership for your preferred circle.