About this role
What you'll do:
- LLM–data interfaces: Research and design how LLMs access and reason over TomTom's data. This covers retrieval strategies for structured and geospatial data, tool and API designs that LLMs can use reliably, text-to-query approaches, and grounding techniques.
- Data representation: Develop representations that make location data usable by AI, such as embeddings for geospatial entities, knowledge graphs and structured context formats.
- Evaluation science: Define how we measure whether LLM-powered systems are right. Build benchmarks, factuality and hallucination metrics, and evaluation datasets for location-grounded tasks, and work with engineers to automate them.
- Model research and development: Design, train and fine-tune ML and deep learning models, including LLMs, computer vision, time-series and graph-based methods, using large-scale, multi-modal data.
- Problem framing: Work with product managers and stakeholders to identify high-impact opportunities and turn them into well-defined scientific problems with clear success metrics.
- Research to production: Write production-quality code and work with AI and software engineers to take solutions from prototype to production, balancing accuracy, latency, cost and scalability.
- State of the art: Keep up with the latest research in LLMs, retrieval and agentic systems, and adapt promising techniques to TomTom's problems.
- Knowledge sharing: Communicate findings clearly to technical and non-technical audiences, and help build a strong scientific culture. External publications and patents are encouraged.
What you'll need:
- Bachelor’s, Master's or PhD in Computer Science, Machine Learning, Statistics, Mathematics, Physics or a related quantitative field, or equivalent professional experience.
- Strong foundations in machine learning, deep learning, statistics and optimization.
- Hands-on experience with modern deep learning frameworks such as PyTorch or JAX, and with the Python scientific stack (NumPy, pandas, scikit-learn).
- Experience with one or more of: LLMs and generative AI, computer vision, time-series forecasting, graph neural networks, or reinforcement learning.
- Proven ability to design experiments, define meaningful metrics and draw sound conclusions from data.
- Experience working with large-scale datasets and distributed computing (e.g., Spark, Databricks, or cloud ML platforms on Azure, AWS or GCP).
- Ability to write clean, maintainable code and to work with engineers toward production deployment.
- Excellent communication skills, with the ability to explain complex technical ideas to diverse audiences.
- A track record of publications at top-tier venues (e.g., NeurIPS, ICML, CVPR, KDD) is a plus.
- Experience with geospatial, mapping, mobility or sensor data is a plus, but not required.