> Markdown version of [/jobs/ext/1999017-data-scientist-ai-search-ranking](https://www.wearedevelopers.com/jobs/ext/1999017-data-scientist-ai-search-ranking). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Data Scientist - AI Search & Ranking - **Company:** trivago GmbH - **Location:** Düsseldorf, Germany (Remote available) - **Experience:** Expert - **Contract:** Permanent contract - **Skills:** Training Data, A/B Testing, Artificial Intelligence, Airflow, Database Queries, E-Business, Python (Programming Language), Machine Learning, Language Modeling, Named Entity Recognition, Recommender Systems, Search Technologies, SQL Databases, Pytorch, Deep Learning, HuggingFace, Build Tools, Machine Learning Operations, Marketplace - **Published:** August 9, 2026 - **Apply:** https://www.careerjet.de/jobad/de7675c3c55e481cd0cd981b031bf98a40 ## About the Role 5+ years building and shipping search, ranking, or recommendation systems in production - Master's or PhD preferred, or equivalent demonstrated expertise. Deep theoretical knowledge and hands-on experience in at least one of: Learning-to-Rank, retrieval and candidate generation, query understanding and NLP, two-tower architectures, or personalisation - and working knowledge of the others. Experience fine-tuning and distilling transformer models for production - building efficient solutions optimised for real-world scale and latency. Solid foundation in experimentation - A/B test design, bias awareness, guardrail metrics, and connecting offline quality to online business outcomes. Strong Python and SQL; hands-on with PyTorch or HuggingFace; familiarity with vector search infrastructure, cloud ML pipelines (Vertex AI, Airflow), and GCP. Clear communicator, entrepreneurial drive, collaborative mentor, and motivation to make progress in ambiguous problem spaces. Outcome-driven: you care about the business impact of your work, not only the sophistication of the model or methodology. Hands-on, outcome driven and analytically rigorous - you take ownership end to end, build with quality in mind, and bring academic or professional rigour into real production environments. A learning and performance mindset - you set ambitious goals, seek feedback, stay curious, and actively explore how AI tools can enhance your work. Nice to have Experience in travel, e-commerce, or two-sided marketplace search. Familiarity with marketplace dynamics and how advertiser signals interact with relevance ranking. Multilingual retrieval experience. ## Description We're building the next generation of search capabilities on top of trivago's existing platform - adding query understanding, semantic retrieval, intelligent ranking, and personalisation to a system that already serves millions of travelers daily. In this role, you will be building a system that handles ambiguous user intent, retrieves the right candidates from millions of hotels, ranks them accurately, and moves measurable business outcomes in production. You will work across three interconnected problem spaces: Query & Intent Understanding: Travelers don't write database queries - they type fragments, use negation, and search for experiences. You will build systems that extract structured meaning from free-text queries across 35+ languages, balance hard constraints with semantic intent, and handle everything from simple city searches to complex multi-constraint natural language queries. Retrieval & Ranking: From candidate generation to neural reranking, you will design systems that balance recall, precision, and latency - close the gap between offline metrics and live conversion, address data bias in behavioural training signals, and personalise results based on in-session and long-term user behaviour. Two-Sided Marketplace: trivago connects travelers with hotels through a marketplace of advertisers. The same hotel appears with different prices from different partners. You will build models that balance user relevance with commercial value - and measure both. Traditional ML, deep learning, and language models all have a place here - your role is knowing when to use each one for real business results. How you'll make an impact: Own components end-to-end - from problem framing through to production deployment and business impact. Build and improve query understanding - intent classification, named entity recognition, slot filling, and semantic interpretation across 35+ languages. Design retrieval systems - candidate generation, dense and hybrid retrieval, and the recall-versus-latency trade-off at query time. Develop ranking and personalisation models - from training data construction and debiasing through to A/B testing and conversion impact - using both short-term in-session signals and long-term user behaviour. Apply and fine-tune language models where they improve the system - result explanation, query rewriting, and agentic approaches for complex and underspecified queries - with clear judgment on latency, cost, and quality trade-offs. Design offline and online evaluation frameworks - relevance judgement pipelines, cold-start evaluation, and retrieval and ranking quality metrics. ## Related Videos - [From Messy Queries to Scalable Systems - How Data Engineering actually works](https://www.wearedevelopers.com/videos/100203-from-messy-queries-to-scalable-systems-how-data-engineering-actually-works) - [Photonic Computing: Programming a New Class of AI Accelerators (incl. Live Coding)](https://www.wearedevelopers.com/videos/100196-photonic-computing-programming-a-new-class-of-ai-accelerators-incl-live-coding) - [100 million days in Vienna: A story of APIs & AI in tourism.](https://www.wearedevelopers.com/videos/93-100-million-days-in-vienna-a-story-of-apis-ai-in-tourism) - [AI Model Management Life Circles: ML Ops For Generative AI Models From Research to Deployment](https://www.wearedevelopers.com/videos/1152-ai-model-management-life-circles-ml-ops-for-generative-ai-models-from-research-to-deployment) - [Beyond Hiring: Building a Sustainable Talent Acquisition Engine at Deutsche Bahn](https://www.wearedevelopers.com/videos/1856-beyond-hiring-building-a-sustainable-talent-acquisition-engine-at-deutsche-bahn) - [Serverless deployment of (large) NLP models ](https://www.wearedevelopers.com/videos/158-serverless-deployment-of-large-nlp-models) ## Related Articles - [The Biggest German Tech Companies](https://www.wearedevelopers.com/magazine/424-the-biggest-german-tech-companies) - [Where to Find German Tech Jobs](https://www.wearedevelopers.com/magazine/366-where-to-find-german-tech-jobs) - [Data Analyst Salary Germany](https://www.wearedevelopers.com/magazine/277-data-analyst-salary-germany) - [Finding Jobs in Germany](https://www.wearedevelopers.com/magazine/375-finding-jobs-in-germany) - [Where To Find Software Engineering Jobs](https://www.wearedevelopers.com/magazine/396-where-to-find-software-engineering-jobs) - [Find a Developer Job: 12 Best Job Sites For Developers](https://www.wearedevelopers.com/magazine/165-find-a-developer-job-12-best-job-sites-for-developers)