Research Scientist, Foundation Model

About the role

Prior Labs is looking for a research scientist to advance foundation models for structured data such as tables, time‑series, and relational datasets. You will design, train, and evaluate novel architectures that can generalise across diverse tasks and datasets, collaborating closely with research engineers to bring ideas to production.

Key responsibilities

• Push the limits of dataset size and context length while improving training and inference efficiency.
• Build multimodal models that combine tabular data with text and other modalities.
• Design architectures and learning methods for time‑series forecasting and anomaly detection.
• Develop relational models that learn across multiple related tables and investigate causal inference capabilities.

Required profile

• Track record of original research contributions demonstrated through publications at leading venues, open‑source work, or deployed methods.
• Strong experience developing and analysing machine‑learning models, including implementation, training, and debugging in PyTorch.
• Solid understanding of deep‑learning training dynamics, generalisation, and common failure modes, with sound experimental judgement.

Required skills

• Python programming.
• PyTorch deep‑learning framework.
• Machine‑learning and deep‑learning model development.