About The Position

NVIDIA is seeking Senior Deep Learning Scientists to advance multimodal AI safety research across product security, content safety, bias, and human-AI interactions. The role focuses on developing methods to align models, reduce harmful behaviors, and apply research insights as practical protections for large language models and multimodal AI products. The scientist will work on safe reasoning, planning, and acting across modalities, with results deployed through Nemotron Omni platforms, guardrails, and external partnerships. The position combines fundamental and applied research with development, training, fine-tuning, deployment, and evaluation of large language architectures and multimodal agentic systems. It also advances post-training and alignment techniques and develops high-quality datasets and benchmarks to assess model accuracy, safety, and task success.

Requirements

  • Master’s degree or equivalent experience, or PhD, in Computer Science, AI, or Applied Math.
  • At least 8 years of relevant work experience.
  • Excellent Python programming skills and strong fundamentals in scalable model development and deep learning frameworks such as PyTorch.
  • Strong knowledge of machine learning and deep learning techniques and modern foundation model architectures, including Transformers and mixture-of-experts models.
  • Hands-on experience post-training multimodal models for multimodal reasoning.
  • Proven ability to manage model development life cycles, including dataset versioning, experiment tracking, and evaluation pipelines.
  • At least 1 year of experience in content safety, ML fairness, AI model security, human-AI interaction, or a related area.

Nice To Haves

  • Strong publication record in top-tier AI and machine learning venues such as NeurIPS, ICML, ICLR, FAccT, or CVPR.
  • Experience applying deep reinforcement learning to train multimodal agents in complex simulation.
  • Proven experience addressing safety and legal or regulatory compliance in multimodal, multilingual, or multi-turn systems.
  • Experience with dynamic and multi-turn alignment benchmarks.
  • Experience in at least one risk area: CBRNE; persuasion and manipulation, including influence operations, disinformation, and erosion of democratic values through AI-generated content; AI-enabled discrimination at scale; mental health risks such as sycophancy and manipulation; LLM security, including backdoors, data poisoning, latent malicious behavior, or structural model vulnerabilities; frontier risks such as model deception, manipulation, and loss-of-control scenarios; or agentic safety involving tool-calling, orchestration, and execution.

Responsibilities

  • Apply fundamental and applied AI safety research to develop, train, fine-tune, and deploy large language architectures for multimodal agentic reasoning systems and guardrails across text, audio, image, video, and tool use.
  • Advance post-training and alignment methods, including instruction tuning, preference optimization, and RLHF/RLVR/MOPD, to improve multimodal agent safety in complex use cases.
  • Research and advance AI safety methods across pre-training, mid-training, post-training, guardrails, and steering techniques.
  • Lead the collection and development of multimodal safety datasets and benchmark them to evaluate model accuracy, safety, and task-completion success.

Benefits

  • Comprehensive benefits package.
  • Equity eligibility.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service