NVIDIA is seeking a Research Scientist in the fields of NLP and deep learning research to contribute to our LLM Research team. We are committed to groundbreaking research that not only pushes the boundaries of knowledge but also makes a significant impact in practical applications. We are particularly interested in exploring methods for alternative language model or multimodal architectures, alternative knowledge distillation and data synthesis methods, alternative scaling laws, information-theoretic approaches, test-time inference algorithms, enhancing factuality, reasoning, long-context, or agentic capabilities of language or multimodal models, and AI safety.
What you’ll be doing:
As a Research Scientist, your role will encompass the following responsibilities:
- Research and propose alternative model architectures that enhance the performance and efficiency of language or multimodal models.
- Investigate novel methods for knowledge distillation and data synthesis for language or multimodal models
- Explore novel methods for test-time inference algorithms
- Analyze multiple novel methods for enhancing factuality, reasoning, long-context, or agentic capabilities of language or multimodal models.
- Research: Undertake research, design, and implementation of novel methods aimed at enhancing the capability of large and small language models.
- Publication: Contribute to the academic community by publishing original research findings in reputable conferences and journals.
- Collaboration: Collaborate closely with fellow team members, as well as interdisciplinary teams within the company, to foster a collaborative research environment.
- Mentorship: Provide mentorship and guidance to interns and students within the research team.
- Technology Transfer: Collaborate with product groups within the university to facilitate the transfer of research technology into practical applications.
What we need to see:
- Completing or recently completed a PhD in CS or equivalent research experiences in research labs with strong publication records at top-tier conferences.
- A strong foundation in the theory and practice of NLP and deep learning.
- Proficiency in all aspects of large language models from training to alignment to inference to evaluation is expected.
- Alternatively, proficiency in all aspects of large multimodal models or diffusion models is expected.
- A track record of outstanding research achievements and contributions.
- Excellent communication skills to effectively convey research findings and collaborate with peers.
