Relational Foundation Model Intern, Modern Data Stack – 2027

Role overview

Contribute to research and engineering on a foundation model that can work across relational databases and heterogeneous graphs. The internship spans machine learning for applications such as forecasting, entity matching, customer retention, and fraud detection, with work potentially covering model design and training through optimization and faster inference.

Responsibilities

- Collaborate with researchers and engineers to improve Transformer- and graph neural network-based models. - Develop approaches that generalize across different relational schemas and graph structures. - Apply machine learning methods to practical prediction, matching, retention, or fraud-detection problems. - Contribute across the model lifecycle, from architecture and training to post-training improvements and inference acceleration.

Requirements

- Be pursuing a master’s or doctoral degree in machine learning, computer science, or a related field. - Have proficiency in Python and deep learning frameworks such as PyTorch. - Bring coursework or research experience developing machine learning algorithms.

Nice to have

- Familiarity with graph-based machine learning. - Research publications at leading machine learning conferences or similar venues.