Principal Scientist, Synthetic Data & Multimodal AI

NVIDIA; New York (NY); Sur place; USD 272 000 - 431 000; Plein temps

Avantages offerts par ce poste

Equity

Benefits

Résumé du poste

NVIDIA seeks a Principal Scientist to set the technical direction for synthetic data generation across its frontier model efforts. You will define and build open-source libraries within the NVIDIA NeMo ecosystem that generate synthetic datasets for text, code, structured, and multimodal data, feeding pre- and post-training of LLMs such as Nemotron.

You will blend hands-on software engineering with applied research, mentor engineers, and collaborate with model teams, external labs, and product

Qualifications

• PhD or equivalent experience in a relevant field.
• 15+ years of engineering and research experience in synthetic data generation, generative modeling, multimodal ML, or related areas.
• Deep technical understanding of LLMs, data shaping pre-training/post-training/RL, and inference frameworks such as vLLM or TGI.
• Proven track record of developing or maintaining software libraries used by a broad developer community.

Responsabilités

• Build and scale data generation pipelines using LLM-based methods with automated quality evaluation.
• Pioneer data generation for agentic and tool-use training including synthetic trajectories and RL components.
• Advance multimodal synthetic data generation across image, document, video, and audio modalities.
• Advance privacy-preserving and safe synthesis using differential privacy and de-identification.
• Develop and maintain open-source libraries and SDKs with clean APIs and documentation.
• Publish original research at top ML/AI conferences and mentor team members.

Connaissances

LLMs

Generative modeling

Multimodal ML

Data pipelines

Publications

Open-source libraries

Formation

PhD in Computer Science / ML or related

Outils

vLLM

TGI

CI/CD tooling

Description du poste

NVIDIA seeks a Principal Scientist to set the technical direction for synthetic data generation across its frontier model efforts. You will define and build open-source libraries within the NVIDIA NeMo ecosystem that generate synthetic datasets for text, code, structured, and multimodal data, feeding pre- and post-training of LLMs such as Nemotron.

You will blend hands-on software engineering with applied research, mentor engineers, and collaborate with model teams, external labs, and product