Principal Research Scientist, Synthetic Data Generation

What you'll do

• Build and scale LLM-based data generation pipelines for reasoning, coding, structured output, and multimodal understanding.
• Pioneer synthetic data generation for agentic behaviors, tool-use training, and reinforcement learning reward modeling.
• Advance multimodal synthetic data generation across image, document, video, and audio formats.
• Develop privacy-preserving and safe synthesis methods including differential privacy and de-identification for regulated domains.
• Build and maintain open-source libraries and SDKs with clean APIs and professional documentation within the NeMo ecosystem.
• Drive software excellence using modern tooling, configuration-based architecture, and robust CI/CD pipelines.
• Publish original research at top machine learning and AI conferences to maintain technical leadership.
• Mentor scientists and engineers to raise the technical bar and grow the next generation of researchers.

What we're looking for

• PhD in Computer Science, Machine Learning, Statistics, or a related field, or equivalent experience.
• 15+ years of engineering and research experience in synthetic data generation, generative modeling, or multimodal machine learning.
• Deep technical understanding of LLMs, pre-training/post-training stages, and inference frameworks like vLLM or TGI.
• Proven track record of developing or maintaining software libraries used by a broad developer community.
• Experience building and optimizing scalable data pipelines for large-scale model training at cluster scale.
• Strong publication record at premier machine learning venues such as NeurIPS, ICML, ICLR, or ACL.
• Experience with multimodal generation, agentic tool-use training, or reinforcement learning post-training.
• Background in privacy-preserving techniques like differential privacy and de-identification for regulated industries.