Senior Scientist: Synthetic Data for LLMs (Open Source)

NVIDIA Corporation; United States; Remote; USD 168,000 - 305,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA Corporation seeks a Senior Scientist to advance synthetic data generation for frontier models, contributing to the NeMo ecosystem to create datasets across text, code, structured, and multimodal data that feed LLM pre- and post-training. You will build pipelines with LLM-based methods, push multimodal data generation (image, document, video, audio), and design open-source SDKs.

Lead research publications, mentor interns, and contribute to a culture of excellence.

Qualifications

• PhD in Computer Science, Machine Learning, Statistics, or related field, or equivalent experience.
• 3+ years of experience in synthetic data generation, generative modeling, multimodal ML, or related areas.
• Strong understanding of LLMs, data shaping for pre- and post-training, and inference frameworks like vLLM or TGI.
• Proven track record of developing or maintaining software libraries used by a broad developer community.
• Excellent publication record at premier venues (NeurIPS, ICML, ICLR, ACL).

Responsibilities

• Build synthetic data generation pipelines using LLM-based methods and automated quality evaluation.
• Advance multimodal synthetic data generation (image, document, video, audio).
• Design and maintain open-source libraries and SDKs with clean APIs and docs.
• Drive software excellence with modern tooling and CI/CD practices.
• Publish original research at top conferences to maintain technical leadership.
• Mentor interns and junior researchers to grow technically.

Skills

LLMs knowledge

Synthetic data generation

Multimodal machine learning

Software libraries

Education

PhD in Computer Science / ML

Tools

Git/CI-CD

Job description

NVIDIA Corporation seeks a Senior Scientist to advance synthetic data generation for frontier models, contributing to the NeMo ecosystem to create datasets across text, code, structured, and multimodal data that feed LLM pre- and post-training. You will build pipelines with LLM-based methods, push multimodal data generation (image, document, video, audio), and design open-source SDKs.

Lead research publications, mentor interns, and contribute to a culture of excellence.