Figma
Support Engineer, AI Infrastructure
About this role
Figma seeks a Support AI Engineer to build AI-powered integrations and automations that enhance product support operations. You'll connect systems like Zendesk and Decagon, implement intelligent workflows, and measure impact on both customer and support team efficiency.
What you'll do
- Design and maintain integrations across support platforms, admin tools, and internal data sources
- Build AI-powered workflows for classification, routing, summarization, and context enrichment
- Bring customer and account metadata into support conversations to enable faster issue resolution
- Establish quality checks, monitoring, and operational safeguards for production AI systems
- Partner cross-functionally with Engineering, Analytics, Security, and Support teams
- Define success metrics and iterate workflows based on adoption and customer outcomes
What they're looking for
- Backend development (Python, Ruby, Go, C++)
- API design and webhook implementation
- LLM and AI workflow experience
- System integration and data orchestration
- SQL/PostgreSQL databases
- Product thinking and stakeholder management
- Monitoring and operational debugging
- AI responsible deployment practices
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Figma
Figma is a design platform that leverages AI and machine learning to power intelligent features like search, ranking, and generative capabilities. The company is hiring software engineers, ML engineers, and support specialists to build scalable infrastructure, AI-powered automations, and support systems that serve millions of users.
- Website
- figma.com
Likely interview questions
- Tell us about a time you built an integration between two complex systems. How did you approach understanding the data flows, and what challenges did you encounter?
- Describe your experience building AI-powered workflows or automations. How did you ensure they were reliable enough for production, and how did you measure their impact?