Senior Machine Learning Engineer, Payments

Airbnb Airbnb · Consumer · United States · Software Engineering

Senior ML Engineer for Payments at Airbnb, focusing on building and deploying LLM-powered workflows, real-time fraud defenses, and personalized checkout flows. The role involves architecting end-to-end solutions, ensuring low latency and high reliability, and mentoring teammates. Requires 5+ years of experience in applied AI/ML, strong programming and data engineering skills, and expertise in modern LLM workflows and scalable inference stacks.

What you'd actually do

  1. Spearhead LLM agents, realtime anomaly detectors, and other breakthrough solutions that solve real-world problems and create product magic.
  2. Collaborate with product, engineering, ops, and data science to spot high leverage opportunities, refine AI/ML requirements, make principled architecture choices, and measure business value with clear, data-driven metrics.
  3. Design, train, deploy, and operate large-scale AI applications for both batch and streaming workloads, ensuring low latency, high reliability, and continuous improvement via automated monitoring and retraining loops.
  4. Mentor and inspire teammates, fostering a collaborative, experimentation-driven environment where cutting edge research meets production excellence and every engineer is empowered to push AI boundaries at Airbnb.

Skills

Required

  • Python
  • Java
  • prompt engineering
  • fine tuning (LoRA, RLHF)
  • hallucination mitigation
  • safety guardrails
  • PyTorch
  • TensorFlow
  • scalable inference stacks
  • vector search
  • orchestration/MLOps platforms
  • large-scale data streaming & processing

Nice to have

  • AI/ML Applications in the Payments domain

What the JD emphasized

  • LLM-powered workflow
  • realtime fraud defenses
  • hyperpersonalized checkout flows
  • latency-first services
  • low latency
  • high reliability
  • continuous improvement
  • rigorous online/offline testing
  • minimize training/inference drift
  • ensure reliable outcomes
  • drift/cost/latency monitoring
  • automated retraining triggers

Other signals

  • LLM-powered workflow
  • realtime fraud defenses
  • hyperpersonalized checkout flows
  • global scale
  • latency-first services