Staff Software Engineer - Instant Access

Plaid Plaid · Fintech · New York, NY · All Departments

Staff Software Engineer on the Instant Access team, building applied AI systems based on LLMs that operate autonomously in production. The role focuses on creating, repairing, and improving financial institution integrations, with responsibilities including defining technical direction for autonomous AI systems, building LLM-powered code generation pipelines, designing evaluation frameworks for non-deterministic systems, and architecting self-healing and multi-layer safety architectures.

What you'd actually do

  1. Define the technical direction and architecture for autonomous AI systems that create, repair, and improve financial institution integrations.
  2. Create novel observability, monitoring, and debugging platforms to provide engineers with deep visibility into the behavior, decisions, and performance of AI-powered systems.
  3. Build LLM-powered code generation pipelines that write, validate, and ship production code end-to-end with minimal human intervention.
  4. Design evaluation frameworks that measure correctness, reliability, and safety for non-deterministic systems operating at scale.
  5. Architect self-healing infrastructure that detects, diagnoses, and repairs itself in under 24 hours.

Skills

Required

  • backend systems
  • platform systems
  • software engineering
  • technical leadership
  • system design
  • observability
  • monitoring
  • debugging
  • LLM
  • code generation
  • evaluation frameworks
  • non-deterministic systems
  • self-healing infrastructure
  • safety architectures
  • risk management
  • trade-off analysis

Nice to have

  • generative AI systems
  • browser automation
  • web automation
  • scraping
  • extraction systems
  • Playwright
  • Stagehand

What the JD emphasized

  • autonomous systems
  • LLM-powered code generation
  • evaluation frameworks
  • non-deterministic systems
  • self-healing infrastructure
  • multi-layer safety architectures
  • hands-on directly in the codebase
  • designing and then building the actual systems you own

Other signals

  • autonomous systems
  • LLM-powered code generation
  • evaluation frameworks for non-deterministic systems
  • self-healing infrastructure
  • multi-layer safety architectures