海角乱伦

Responsibilities :

• Lead design and implementation of the context and orchestration layer; set engineering standards for reliability, latency, and cost.
• Partner daily with Product and GTM teams to convert messy GTM data into shipped, impactful outcomes.
• Create the collective memory: Ingest and unify multi-source data into a multi-level context graph with strong tenant isolation.
• Orchestrate agentic systems: Design planner/executor patterns, tools, and policies (including MCP-style interfaces) that convert context into content and into actions; define simple evaluation harnesses.
• Deliver natively: Expose capabilities via apps, chat, and integrations to reduce user context switches.
• Prove outcomes: Define success metrics (tasks auto-completed, adoption/retention, pipeline lift), and wire observability for rapid ship-learn-iterate cycles while meeting latency targets.
• Balance cost and reliability: Tune for accuracy, latency, and cost for retrieval and agents; implement fallbacks and safeguards for real-world load.
• Raise the bar: Write RFCs, lead design reviews, mentor peers, and improve code quality, SLOs, and on-call practices.

Qualifications :

• 4 years building and owning backend or platform systems end-to-end, with measurable business impact.
• Strong Python skills; capable of deploying AWS cloud infrastructure and running production services.
• Experience integrating messy, multi-source data into models that products can reason over, with strong privacy, reliability, and multi-tenant instincts.
• Comfortable making decisions with roughly 70% of information, instrumenting what matters, and iterating quickly.
• Nice to have: Agent orchestration/planning, retrieval or graph-shaped context, evaluation frameworks, distributed systems at scale.

Hybrid availability: Able to work 3 days per week in San Francisco, New York City, or Vancouver (or willing to relocate on an agreed timeline)