Senior AI Researcher - Pre-training Data (m/f/d)
Core
Designing scientific methodology and building systems for foundation model pre-training data, focusing on data quality estimation, synthetic data generation, and curriculum learning.
Role type
Senior AI Researcher (Data-Centric AI)
Builds
Scalable data pipelines and novel algorithms for data curation and quality scoring
Domain
Large-scale language modeling, data-centric AI, foundation models
Deliverable
production ML models
Required skills
Machine learning theory, experiment design, statistical methods, information-theoretic reasoning, Python, PyTorch
Preferred skills
PhD in ML/NLP, publications in top-tier venues, experience training foundation models from scratch
Technologies
PyTorch
Responsibilities
Design targeted ablations across various scales, develop novel algorithms for estimating data quality, research synthetic data generation and curriculum learning, partner with engineers to scale research pipelines
Seniority
Senior, hands-on IC