Data Scientist - AI Search & Ranking

As a Senior Data Scientist in our Search & Ranking team, you will shape how millions of travelers find and compare hotels on our platform. Our team powers hotel discovery across 35+ languages and 55 countries, processing millions of searches daily. We're building the next generation of AI-powered search that combines personalized ranking with cutting-edge semantic understanding—and we need your expertise to take it further.

You'll work at the intersection of search ranking, natural language processing, and AI to improve query understanding, relevance, and personalization. With a focus on user-intent and AI‑powered search initiatives, you'll enhance semantic retrieval, model user intent, and refine personalization to better match travelers with their ideal stays. Working alongside talented data scientists, ML engineers, and product teams, you'll design evaluation frameworks, run large-scale experiments, and influence product decisions that increase user satisfaction and business outcomes. If you're passionate about solving complex search problems and making an impact at global scale, we'd love to hear from you!

How you'll make an impact:

• Design and develop scalable search and ranking architectures that integrate semantic retrieval, query/intent understanding, deep learning and Learning‑to‑Rank algorithms
• Build and productionize NLP/ML components for real-time query understanding and personalized ranking with low latency and high reliability
• Fine‑tune and deploy Large Language Models (LLMs) and build agentic AI workflows for query understanding, intent classification, and semantic retrieval
• Develop ranking algorithms that balance user relevance, semantic understanding, and marketplace dynamics (pricing, advertiser signals, conversion) to optimize both user satisfaction and platform value
• Create robust evaluation frameworks and metrics (e.g., A/B testing, NDCG, MRR, ) to continuously improve relevance and personalization
• Apply causal inference, sequential testing, and advanced statistical methods to optimize search quality and user engagement
• Collaborate cross‑functionally with engineers, product managers, and fellow data scientists to ship production ML systems that impact millions
  

What you'll need to thrive:

• 5+ years in search, information retrieval, personalized ranking systems, NLP, machine learning, or a deep learning with proven product impact
• Deep expertise in semantic search: query understanding (intent classification, Named Entity Recognition), embedding‑based retrieval with vector databases (e.g. qdrant, vespa), and Learning‑to‑Rank.
• Proficiency with LLMs and transformer architectures (BERT, T5, GPT): fine‑tuning, prompt engineering, and few‑shot learning for search applications
• Experience building production ML systems with real-time inference, model serving, and monitoring at scale
• Strong Python and SQL skills; proficiency with ML frameworks (PyTorch, TensorFlow, HuggingFace)
• Familiarity with big data and cloud ecosystems (e.g., GCP, BigQuery), MLOps tools (e.g., Airflow, Kubeflow), containerization and feature stores
• Solid statistics foundation with hands‑on experience designing and analyzing large‑scale A/B tests
• Strong problem‑solving skills with entrepreneurial passion: comfortable navigating ambiguity, balancing rapid prototyping with rigorous research, and adapting solutions to meet both immediate and long‑term business needs
• Excellent communication and collaboration skills: proven ability to work in cross‑functional teams, explain complex ideas to non-technical stakeholders, and mentor junior data scientists
• Master's or PhD in NLP, Computer Science, Machine Learning, Statistics, Mathematics, or related quantitative field
  Stand out with:
• Experience with LLM orchestration and building agentic workflows (e.g., LangChain, LlamaIndex) for semantic search applications
• Experience with two‑sided marketplace ranking and systems in e‑commerce, travel, or online advertising
• Experience optimizing AI systems in production with attention to latency, cost, and response quality