Research Scientist

• Design and evaluate algorithms that turn noisy, incomplete probe data into reliable speed and closure information
• Prototype in Python, then implement in Java in production services and own the change through rollout
• Develop against ground truth and use the results to shape future development
• Evolve existing traffic services used by millions of drivers
• Use AI coding assistants such as Claude Code for exploration, implementation, and analysis, while reviewing outputs and retaining full accountability
• Collaboratively run production systems, including monitoring, incident management, and production support
• Improve traffic services processing 5–6 billion kilometers of driving data every day across 70+ countries

Applied Scientist

• Research and design how LLMs access and reason over TomTom's structured and geospatial data
• Develop retrieval strategies, tool and API designs, text-to-query approaches, and grounding techniques
• Develop AI-ready representations of location data, including geospatial embeddings, knowledge graphs and structured context formats
• Define benchmarks, factuality and hallucination metrics, and evaluation datasets for location-grounded tasks
• Design, train and fine-tune ML and deep learning models, including LLMs, computer vision, time-series and graph-based methods
• Work with product managers and stakeholders to frame high-impact opportunities as scientific problems with clear success metrics
• Write production-quality code and collaborate with AI and software engineers to move solutions from prototype to production
• Balance accuracy, latency, cost and scalability
• Keep up with research in LLMs, retrieval and agentic systems and adapt techniques to TomTom's problems
• Communicate findings to technical and non-technical audiences and contribute to scientific culture