Principal Demo Engineer

Intercom Intercom · Enterprise · San Francisco, CA · Customer Success & Solutions

The Principal Demo Engineer will own and build the platform and tooling for Intercom's AI Customer Agent (Fin) demos. This involves creating reliable, fast, and safe demo environments, seeded data, integrations, and reusable components for Solutions Engineers. The role also involves embedding with product teams to support new releases, create training materials, and collaborate across Product, GTM, and Engineering.

What you'd actually do

  1. Own the Fin demo platform end-to-end. Multi-tenant mock workspaces, seeded data, configuration, reset infrastructure, and guardrails that make demos reliable, fast, and safe.
  2. Build first-class demo features into Fin itself. Create UI and workflows so SEs can spin up, reset, and tailor demo accounts in minutes - not hours.
  3. Use AI to prepare demos. Use Fin itself (and other models) to research prospects, synthesize realistic org structures and conversations, and auto-configure demos for specific verticals and use cases.
  4. Let SEs move faster. Build templates, libraries, and scripting hooks so SEs can compose complex demo scenarios with guardrails - not raw access to production.
  5. Show Fin across every surface it touches. Web Messenger, mobile SDKs, voice, email, Slack, API-driven workflows. The demo should match how customers will actually deploy.

Skills

Required

  • TypeScript
  • Ruby
  • Go
  • Python
  • web development
  • mobile development
  • SDKs
  • APIs
  • RESTful APIs
  • SaaS integrations
  • LLM product experience
  • RAG
  • tool use
  • agents
  • evals
  • design sensibility
  • customer support experience
  • contact center tooling
  • CX tooling

Nice to have

  • Experience building internal platforms or developer tooling at scale.
  • Experience designing or maintaining multi-tenant demo or sandbox environments

What the JD emphasized

  • 6+ years building software in a customer-facing technical role
  • Hands-on AI/LLM product experience
  • You've built with LLMs, you understand RAG, tool use, agents, and evals, and you know the practical limits of current models.
  • You've made yourself measurably more productive with Claude Code or equivalent agentic coding tools.
  • You can show concrete examples of where AI doubled your output

Other signals

  • AI Agent
  • customer support
  • demo platform
  • internal tooling