AI Research Scientist

NVIDIA is searching for outstanding AI researchers to join the NVIDIA Research Korea team. Our mission is to advance the frontiers of foundation models and agentic AI through data-centric research. We focus on developing more capable, efficient, and self-improving AI systems by improving how models learn from data, spanning pre-training, post-training, and inference-time scaling. Our research explores synthetic data generation, data selection and curation, data-efficient learning, agentic reasoning, and iterative self-improvement.

A broader goal of our research is to advance Sovereign AI by developing generalizable methods and reproducible recipes that enable countries and language communities to build advanced AI systems without relying on massive amounts of human-annotated data. Building on NVIDIA Nemotron open models, we aim to make cutting-edge foundation models and agentic AI capabilities more accessible across languages, regions, and applications.

You will join a highly collaborative research team tackling fundamental challenges in foundation models and agentic AI. Our team aims to conduct groundbreaking research, contribute to the open-source AI ecosystem, and translate research innovations into NVIDIA technologies and products.

What you'll be doing:

• Conduct cutting-edge research in foundation models and agentic AI, including reasoning, planning, tool use, and autonomous problem-solving.
• Develop novel methods for data-efficient learning, data curation, and scalable model training for both pre-training and post-training.
• Explore self-improvement techniques, including RL, distillation, self-training, and learning from model-generated feedback.
• Advance the capabilities and efficiency of foundation models and agentic systems through innovative pre-training, post-training, and inference-time methods.
• Develop advanced Korean-language foundation models and agentic systems, establishing methods and training recipes that generalize to other languages and regions.
• Publish original research at leading conferences and contribute to open-source research.
• Collaborate with researchers across NVIDIA and external academic institutions.
• Mentor research interns from the NVIDIA-KAIST Joint AI Research Lab and contribute to a collaborative research environment with KAIST.
• Present research findings at conferences, workshops, and other scientific events.
• Transfer research innovations to NVIDIA product teams.

What we need to see:

• PhD in AI, Computer Science, Electrical Engineering, or a related field (or equivalent experience).
• At least 4 years of relevant research experience.
• Strong expertise in one or more of the following areas: LLMs, pre-training, agentic AI, reasoning, reinforcement learning, efficient model training, synthetic data generation, data-efficient learning, distillation, or self-improving AI.
• Publications at leading conferences such as NeurIPS, ICML, ICLR, COLM, ACL, EMNLP, or CVPR.
• Strong understanding of modern deep learning methods and foundation models.
• Excellent programming skills in Python and experience with deep learning frameworks such as PyTorch.
• Ability to identify important research problems, develop novel approaches, and conduct rigorous empirical evaluations.
• Excellent communication skills and ability to collaborate effectively across research teams in English.

Ways to stand out from the crowd:

• Research contributions to state-of-the-art foundation models, agentic systems, reasoning models, or autonomous AI.
• Experience with large-scale distributed pre-training, efficient post-training, or inference-time scaling.
• Impactful contributions to Korean-language AI, multilingual foundation models, or cross-lingual transfer.
• Significant open-source contributions to foundation models, datasets, or agentic AI frameworks.

NVIDIA is widely considered to be one of the technology world's most desirable employers. We have some of the most forward-thinking and productive people in the world working for us. Looking forward to hear from you!