Research Scientist, Data

You will work on the most important aspect of Scientific AI creation: evaluations and data. This means constructing cutting-edge evaluations based on advanced scientific use cases, sourcing and procuring external datasets, integrating internally generated experimental data into the training stack, constructing training environments for RL. You’ll ensure that the team always has the right assets, in the right shape, to evaluate and improve AI models.

You will work with computational and experimental scientists to translate complex scientific workflows into rigorous evaluations and agentic benchmarks, and partner with pretraining, midtraining, and reinforcement learning researchers to identify the data models needed, then build the datasets, environments, and pipelines to deliver it. Your goal will be to create a tight feedback loop between scientific use cases, model evaluation, and training data.