AI Platform Engineer - Scale & Production

Jobtailor; California (MO); On-site; USD 120,000 - 180,000; Full time

Job summary

Notion is seeking an engineer to own the prototyping, development and scaling of AI platform primitives. You will collaborate with product teams to provide production-ready guardrails that help new AI features ship faster with less duplicated work.

You’ll operate critical AI systems in production, using observability and diagnostics to understand provider/model behavior, debug failures, improve latency and cost, and evolve systems with minimal user disruption.

Qualifications

• Experience with AI systems at scale and shared platforms.
• Experience on AI product development including prompting, evals, model integrations, or quality measurement.
• Built and scaled data pipelines (Spark or Ray).
• Full-stack experience in TypeScript and Node.js ecosystem.
• Experience building MLOps and ML serving infrastructure.

Responsibilities

• Own prototyping, development and scaling of AI platform primitives.
• Partner with product teams to provide production-ready guardrails for faster feature shipping.
• Work across infrastructure, libraries, APIs, and product integration points to enable easy adoption of AI capabilities.
• Operate critical AI systems in production, using observability to understand behavior, diagnose issues, and optimize latency and cost.
• Help Notion safely adopt new models and providers through versioning and controlled rollouts.

Skills

AI systems

LLMs

ML platform

Data pipelines

Full-stack TypeScript/Node.js

MLOps

Observability

Tools

Apache Spark

Ray

TypeScript

Node.js

Job description

Notion is seeking an engineer to own the prototyping, development and scaling of AI platform primitives. You will collaborate with product teams to provide production-ready guardrails that help new AI features ship faster with less duplicated work.

You’ll operate critical AI systems in production, using observability and diagnostics to understand provider/model behavior, debug failures, improve latency and cost, and evolve systems with minimal user disruption.