Software Engineer, Code Platform

Figma Figma · Enterprise · Canada +1 · Engineering

Software Engineer on Figma's Code Platform team, focusing on building AI-forward systems that translate between design models and LLM-friendly representations. The role involves designing and implementing AI-powered and agentic workflows for code generation and design interpretation, including building evaluation frameworks for AI outputs and optimizing performance.

What you'd actually do

  1. Design, build, and improve code ↔ design translation pipelines, including AI-powered and agentic workflows
  2. Build and iterate on LLM integrations that power code generation, design interpretation, and agentic feature surfaces
  3. Design evaluation frameworks for AI-generated code outputs quality, correctness, and format fidelity
  4. Diagnose and resolve performance bottlenecks and long-running tasks that impact product SLAs
  5. Improve correctness, reconciliation, and fidelity across complex design system and layout edge cases

Skills

Required

  • 5+ years of software engineering experience, especially in web, platform, or infrastructure engineering
  • Strong TypeScript/JavaScript and modern frontend fundamentals
  • Deep experience with declarative UI systems (React, JSX, or similar)
  • Familiarity with ASTs, code transformation, or compilation concepts
  • Experience building features that integrate LLMs or AI models into a product not just using AI tools, but shipping with them
  • Experience debugging performance, rendering, or pipeline latency issues
  • Experience building complex, cross-team platforms with multiple stakeholder groups

Nice to have

  • Experience with prompt engineering, RAG patterns, or model evaluation
  • Experience with code generation models or AST-level code manipulation
  • Prior work on agentic workflows or multi-step LLM pipelines
  • Experience with MCP, tool-use APIs, or AI developer ecosystems
  • Experience building design systems or component libraries
  • Platform or infrastructure engineering experience (APIs, observability, SLAs)
  • Rendering, layout, or graphics pipeline knowledge

What the JD emphasized

  • AI-powered and agentic workflows
  • LLM integrations
  • agentic feature surfaces
  • AI-generated code outputs quality, correctness, and format fidelity
  • evaluating AI-generated code outputs
  • integrating LLMs or AI models into a product

Other signals

  • integrating AI deeply into how developers experience Figma
  • AI-powered and agentic workflows
  • LLM integrations that power code generation, design interpretation, and agentic feature surfaces
  • AI-generated code outputs quality, correctness, and format fidelity