Senior Staff Machine Learning Engineer, Post Training

Airbnb Airbnb · Consumer · United States · Software Engineering

Senior Staff Machine Learning Engineer focused on post-training LLM optimization and productionization for customer support AI initiatives at Airbnb. The role involves fine-tuning, alignment, RAG/Search, evaluation, and feedback loops to enhance customer experience through AI.

What you'd actually do

  1. Work with large scale structured and unstructured data; explore, experiment, build and continuously improve foundation models for Airbnb product, business and operational use cases.
  2. Create a multi-year tech roadmap that enables our team to stay on the leading edge of the rapidly evolving AI landscape and leverage the best in class technologies to deliver customer benefits.
  3. Continuously evaluate recent and upcoming large foundational models, ensuring the selection and refinement of the highest quality models for enhanced performance and efficiency.
  4. Hands-on prototype, develop and productionize LLM models and pipelines at scale, including both batch and real-time use cases.
  5. Drive key AI architectural decisions for products, and contribute to Airbnb’s ML platform architecture and strategy.

Skills

Required

  • PhD in Computer Science, Machine Learning, Mathematics, Statistics, or related technical field.
  • 10+ years of experience with developing machine learning models and products at scale from inception to business impact.
  • Programming experience in Python
  • hands-on experience with frameworks such as PyTorch
  • Proven record of training, fine tuning, optimizing models and inference run-time
  • Post-training experience in areas like data processing for fine-tuning; responsible LLMs; LLM alignment; reinforcement learning; efficient training and inference; language model evaluation; and/or multilingual and multimodal modeling.
  • Or specialized experience in runtime optimizations, model quantization, compression, on-device inference, GPU inference, pytorch, kernel development

Nice to have

  • Responsible LLMs
  • LLM alignment
  • reinforcement learning
  • efficient training and inference
  • language model evaluation
  • multilingual and multimodal modeling
  • runtime optimizations
  • model quantization
  • compression
  • on-device inference
  • GPU inference
  • pytorch
  • kernel development

What the JD emphasized

  • Post-training experience
  • training, fine tuning, optimizing models and inference run-time

Other signals

  • Generative AI
  • LLM fine-tuning
  • optimization
  • productionize LLM models