Machine Learning Scientist II - Personalization

Expedia Expedia · Hospitality · Geneva, Switzerland

Machine Learning Scientist II at Expedia Group focused on building the deep learning foundations for a centralized, real-time personalization engine. The role involves developing, evaluating, and refining ML models for prediction, ranking, and recommendation, designing end-to-end ML workflows, and collaborating with cross-functional teams to integrate ML solutions into production systems. It also includes analyzing large datasets, applying ML principles, and demonstrating familiarity with AI-driven systems and their safe integration into real-world products. The role spans neural recommendation, ranking, representation learning, sequential learning, and emerging applications like agent personalization and memory.

What you'd actually do

  1. Develop, evaluate, and refine machine learning models to solve clearly defined business problems, including tasks such as prediction, ranking, recommendation, and optimization for Expedia Group products and services.
  2. Design and implement end-to-end ML workflows, including data exploration, feature engineering, model training, validation, and A/B experimentation, ensuring robust measurement of impact and model performance.
  3. Collaborate with software engineers, data scientists, product managers, and other stakeholders to integrate ML solutions into production systems, focusing on reliability, scalability, and maintainability.
  4. Analyze large-scale, real-world datasets to uncover insights, identify opportunities for model and feature improvements, and communicate findings and tradeoffs to both technical and non-technical audiences.
  5. Apply sound statistical and machine learning principles, including model evaluation, error analysis, and bias/variance tradeoffs, to improve model quality and ensure responsible use of data and algorithms.

Skills

Required

  • Python
  • deep learning frameworks such as PyTorch, TensorFlow, Keras, or JAX
  • data processing tools
  • designing, implementing, and evaluating deep learning models
  • end-to-end ML pipelines or services
  • applying machine learning or statistical modeling to real-world problems
  • building, evaluating, and deploying models to production environments

Nice to have

  • Advanced degree (Master’s or PhD) in Machine Learning, Computer Science, Statistics, or a related quantitative discipline with a focus on applied machine learning, deep learning, or statistical modeling.
  • taking deep learning models from concept through experimentation and iteration, with exposure to deployment, monitoring, or scaled adoption in production systems.
  • neural recommendation, ranking, retrieval, two-tower models, or learned embeddings and representations for travelers, items, sequences, or other structured entities.
  • sequential models, transformers, or attention-based architectures, including model design and evaluation beyond applying a standard pretrained-model fine-tuning recipe.
  • implementing recent research, academic publications, substantial research projects, open-source contributions, internships, or relevant competitions.
  • large-scale data processing, ETLs, training pipelines, A/B experimentation, model serving, monitoring, or other production ML practices.
  • agent personalization, learned memory, LLM evaluation, or hybrid LLM-recommender systems grounded in rigorous modeling and experimentation rather than only out-of-the-box APIs or retrieval-augmented generation workflows.

What the JD emphasized

  • deep learning foundations
  • neural recommendation and ranking
  • agent personalization
  • memory
  • evaluation
  • end-to-end ML workflows
  • integrate ML solutions into production systems
  • large-scale, real-world datasets
  • AI-driven systems
  • applying AI/ML concepts to real world products
  • safely integrating and operating AI/ML‑enabled solutions

Other signals

  • centralized, real-time personalization engine
  • deep learning foundations
  • neural recommendation and ranking
  • agent personalization
  • memory
  • evaluation
  • end-to-end ML workflows
  • integrate ML solutions into production systems
  • large-scale, real-world datasets
  • AI-driven systems
  • applying AI/ML concepts to real world products
  • safely integrating and operating AI/ML‑enabled solutions