Senior Data Scientist

About the role

Join the Search Ranking team within our Global Search tribe to refine search functionality across 60 countries and 35 languages. You will manage deep learning ranking models that process over 80 million daily searches, focusing on autonomous experimentation and production-grade system performance. In this capacity, you will own the design and continuous optimization of neural architectures that directly influence how users discover food and services. You will act as a technical leader in defining experiment strategies that balance scientific rigor with rapid business impact. The role requires you to translate ambiguous product questions into robust modeling solutions that scale reliably. You will be responsible for ensuring that insights derived from data are translated into actionable improvements in search relevance. Ultimately, your work will shape the efficiency and accuracy of one of the core surfaces connecting customers with merchants across the globe.

Key facts

What you'll do

• Design, build, and deploy deep neural ranking architectures including DCN-V2, MMoE, and Two-Tower systems to handle complex query and candidate interactions.
• Manage the full model lifecycle from initial offline evaluation through online A/B testing, ensuring robust monitoring for model drift, latency issues, and pipeline integrity in production.
• Utilize LLM coding agents like Claude Code and Gemini to automate feature engineering, streamline data preprocessing, and accelerate experiment pipelines.
• Review large volumes of autonomous experiment runs to identify subtle performance signals and systematically optimize Learning to Rank architectures.
• Partner closely with backend and data engineers to align model performance with business goals, ensuring that technical implementations support key commercial objectives.
• Mentor junior team members by providing code reviews, technical guidance, and clear documentation of modeling approaches and best practices.
• Establish and maintain model registries and experiment tracking frameworks to ensure reproducibility and transparency across the modeling lifecycle.
• Implement CI/CD practices for machine learning, enabling frequent and reliable deployment of new ranking models with minimal manual intervention.
• Conduct rigorous offline and online analysis to validate that changes to ranking models lead to meaningful improvements in user satisfaction and engagement.
• Collaborate with product managers to define success metrics and establish evaluation frameworks that measure the impact of search improvements.
• Explore embedding-based retrieval techniques to enhance the efficiency and relevance of candidate generation within the search pipeline.
• Contribute to the broader data science community within Delivery Hero by sharing insights, methodologies, and tools that improve modeling standards.

Requirements

• Hold a Master degree or a Bachelor degree with at least 6 years of relevant experience in Computer Science, Mathematics, Physics, or a closely related quantitative field.
• Possess a minimum of 4 years of industry experience as a Data Scientist or Machine Learning Engineer focused on high-traffic production environments where model performance directly impacts revenue.
• Demonstrate proven expertise in deep learning ranking architectures and Learning to Rank methods, including a clear understanding of pointwise, pairwise, and listwise approaches.
• Show proficiency with LLM coding agents for autonomous development, using these tools to prototype architectures and accelerate implementation tasks.
• Exhibit strong command of Python, PyTorch, TensorFlow, scikit-learn, SQL, and big-data processing tools such as PySpark or Scala in real-world projects.
• Have hands-on experience with MLOps practices, including CI/CD pipelines, Metaflow, model registries, and experiment tracking systems to ensure reliable model operations.
• Bring experience working with large-scale data where performance, latency, and reliability constraints are critical considerations.
• Show a strong sense of ownership for model performance and reliability, with the ability to diagnose issues in production systems quickly.

Nice to have

• Experience with embedding-based retrieval and multi-task learning to further enhance the quality and efficiency of search operations.
• Familiarity with cloud environments such as AWS or GCP, including services related to compute, storage, and machine learning platforms.

Skills & tools

• Python, PyTorch, TensorFlow, scikit-learn, SQL, dbt, PySpark, Scala.
• DCN-V2, MMoE, Two-Tower architectures.
• LLM coding agents (Claude Code, Gemini).
• MLOps (Metaflow, CI/CD, model registry).
• Cloud platforms (GCP, AWS).

Practical notes

• Hybrid work model requires 2 days per week at the Berlin campus.
• Benefits include 27 days of holiday, 1,000 Euro educational budget, corporate pension plan, and employee share purchase plan.
• Additional perks: digital meal vouchers, gym access, meditation, and life/accident insurance.
• Relocation support is available for international candidates.
• We provide reasonable accommodations for candidates; contact inclusion@deliveryhero.com for support.
• Severely disabled applicants with equal qualifications receive preferential consideration.

About the company

Glovo is part of the Delivery Hero Group, the world's pioneering local delivery platform, our mission is to deliver an amazing experience - fast, easy, and to your door. We operate in around 65 countries worldwide. Headquartered in Berlin, Germany.