Principal Machine Learning Scientist

Overview

As Principal Machine Learning Scientist, you anchor the Trips discovery engine, shaping ML strategy and production systems that power search, discovery, and itineraries for millions of users. You bridge cutting-edge AI research with scalable engineering, driving multi-objective business impact such as engagement and conversions. You lead advanced models for ranking, retrieval, and recommendations, while mentoring a high-performing team. This role combines technical leadership with hands-on delivery in a fast-paced, graph- and sequence-driven travel domain.

Pay / Benefits

• 远程工作友好
• 灵活安排
• 旅行福利与折扣
• 捐赠匹配
• 学费资助
• 健康福利

Responsibilities

• Drive technical roadmap for Search, Retrieval, Ranking, and Recommendations within the Trips vertical
• Translate business goals into scalable ML architectures and production systems
• Design and scale next-generation ranking and recommendation models
• Oversee low-latency, high-throughput pipelines (multi-stage retrieval, vector search) for real-time processing
• Mentor and coach senior and mid-level ML scientists; promote MLOps, A/B testing, data privacy, and code quality
• Collaborate with Product Managers, Engineering Leads, and Data Science peers to optimize multi-task objectives
• Act as the primary technical authority for ML initiatives in the Trips vertical

Key requirements

• 8+ years of industry experience deploying large-scale ML models in production
• Proven track record of shipping systems handling millions of active users
• Deep expertise in retrieval and ranking, particularly multi-task optimization
• Hands-on experience with sequential/temporal modeling for real-time user dynamics
• Strong representation learning and multi-modal embedding techniques
• Proficiency in Python and DL frameworks (TensorFlow, PyTorch)
• Experience with distributed computing (Spark, Ray) and cloud (AWS/GCP)
• Familiarity with graph concepts and potentially GNNs is a plus
• Experience in travel tech or e-commerce and handling constrained inventories would be beneficial

• 领导力 and mentoring
• 跨职能协作与沟通
• 结果导向与数据驱动
• Multi-Task Learning (MTL) / MMoe
• Sequential & Temporal Modeling
• Graph Neural Networks (GNNs) / knowledge graphs

…