Senior Research Scientist, Fundamental LLM Research for Knowledge, Reasoning, and Agents

NVIDIA•Us, CA

10d

About The Position

NVIDIA is searching for an outstanding researcher working on Large Language Model (LLM) research team. We are passionate about research that pushes boundaries but also has impact in the real world. We are particularly excited about methods for post-training and alignment, principled approaches to synthetic data generation and filtering, advanced reasoning and inference algorithms for LLMs, novel learning paradigms and LLM architectures, and scientific understanding about the fundamental limits and capabilities of LLMs. You will work within an amazing and collaborative research team that consistently publishes at the top venues in machine learning and natural language processing fields. Our existing expertise includes deep learning, NLP, computer vision. Your contributions have the chance to create real impact on our products.

Requirements

PhD in Computer Science or Computer Engineering (or equivalent experience).
At least 6 years of research experience (demonstrated by publication records spanning across 5+ years) in artificial intelligence, machine learning, natural language processing, computer vision or related subjects
A history of research success exemplified by a strong publication record and awards.
Excellent knowledge of theory and practice of deep learning and natural language processing.
Background in LLM training, alignment, and evaluation is expected.
Excellent programming skills in Python and PyTorch.
Hands-on experience with large-scale model training including data preparation and model parallelization (tensor and pipeline) is required.
Excellent communications skills.

Responsibilities

Explore alternative avenues to unlock new capabilities in language models, including advanced knowledge acquisition techniques and innovative learning and decoding algorithms.
Innovate new learning paradigms that incorporate agency into the training of language models, such as enabling self-reflection and targeted knowledge enhancement.
Enable learning from multi-modalities beyond written text, such as acquiring physical commonsense knowledge through interactions with real-world environments.
Publish original research.
Collaborate with other team members and teams.
Mentor interns.
Speak at conferences and events.
Work with product groups to transfer technology.
Collaborate with external researchers.