- Level
- Senior
- Experience
- 8+ years
- Work style
- Remote
- Full-time
- Area
- Security engineering
At a glance
The role in plain English
This role focuses on advancing the security and alignment of large language models and multimodal AI systems. It suits experienced deep learning scientists with strong programming skills and a background in AI security research. Day-to-day work involves developing, training, and evaluating models, researching security techniques, and collaborating with cross-functional teams to scale security solutions for AI agents.
What you'll do
- Develop and deploy secure large language models for multimodal agentic reasoning
- Advance post-training and alignment methods to improve model security
- Research and implement techniques for model and system security
- Lead collection and benchmarking of multimodal security datasets
- Collaborate with engineers and research teams to scale LLM security solutions
What they're looking for
Essential
- Master’s or PhD in Computer Science, AI, or Applied Math
- 8+ years relevant work experience
- Excellent Python programming skills
- Experience with deep learning frameworks like PyTorch
- 1+ years in LLM security, frontier risks, agentic safety, or multi-turn evaluation
Nice to have
- Publications in top-tier AI/ML venues
- Open-source AI safety contributions
- Experience with multilingual AI security
- Experience with auto red-teaming
Core skills
- Python
- PyTorch
- Deep learning
- LLM security
- Multimodal models
- Model alignment
- Dataset benchmarking
- AI safety
The shape of the job
Technical depth
5/5
Requires deep knowledge of model architectures, security, and advanced alignment techniques.
Coding required
4/5
Excellent Python skills and scalable model development are essential.
Experience required
8+ years
The listing asks for at least 8 years of experience.
Hands-on focus
5/5
Hands-on development, training, and evaluation of models required.
Cyberend's reading of the listing's duties and requirements, not the employer's own scores.
Application toolkit
Make your application count
See which requirements your CV meets and what is missing.
Your CV is used for this answer only and never stored.
Questions to prepare for
- 01Describe your experience securing large language models against adversarial manipulation.
- 02How have you applied post-training alignment methods to improve AI safety?
- 03What techniques do you use to benchmark the security of multimodal models?
- 04Explain your approach to detecting and mitigating model vulnerabilities such as backdoors or data poisoning.
About NVIDIA
NVIDIA is a technology company known for GPU-accelerated computing and AI research.
Full job descriptionAs published by NVIDIA
NVIDIA helped ignite the modern AI era through GPU accelerated deep learning. Now, we’re helping shape what comes next—and the security that makes it possible. AI will do more than respond to questions. It will use tools, navigate sophisticated environments, and take on greater responsibility in the world. Securing these capabilities is one of the defining technical challenges of our time. Join us to turn scientific discovery into security breakthroughs that help people build and use AI with greater confidence.
We’re looking for Deep Learning Scientist to join our AI Security and Responsibility efforts. You’ll advance security alignment for large language models and develop guardrails that help the systems they power resist adversarial manipulation and operate within trusted boundaries. NVIDIA’s work creates opportunities to carry new ideas from experiments into practice. If you’re driven to explore how powerful systems can fail and determined to make them stronger, bring that curiosity and persistence to Nvidia.
What you will be doing:
- Apply fundamental and applied security research to develop, train, fine-tune, and deploy large language models for multimodal agentic reasoning systems and guardrails encompassing text, audio, image, video, and tool usage.
- Advance post-training and alignment methods including instruction tuning, preference optimization, and RLHF/RLVR/MOPD to improve security of multimodal agents for complex use cases.
- Model and System Security: Research and deploy techniques beyond post-training, such as instruction hierarchy, risk detection, and new cyber gyms.
- Lead the collection, development, and benchmarking of multimodal security datasets, ensuring high-quality evaluation of model accuracy, safety, and task completion success.
- Partner with engineers, data scientists, research teams, and product managers across NVIDIA to scale solutions for LLM security and agentic safety.
What we need to see:
- Master’s degree (or equivalent experience) or PhD in Computer Science, AI, or Applied Math with 8+ years of relevant work experience.
- Excellent programming skills in Python with strong fundamentals in scalable model development and deep learning frameworks like PyTorch.
- Strong knowledge of ML/DL techniques and modern foundation model architectures, including Transformers and mixture-of-experts models.
- Core Safety Expertise: 1+ years of dedicated experience or research in at least one of the following areas:
- LLM Security: Focus on backdoors, data poisoning, latent malicious behavior, and structural model vulnerabilities.
- Frontier Risks: advanced alignment challenges, including model deception, manipulation, and loss-of-control scenarios.
- Agentic Safety: Focus on LLM-level security for autonomous systems, including multi-turn tool-calling, orchestration, and execution risks.
- Multi-turn Security Evaluation: Focus on robust, scalable automated evaluation methodologies for conversational and iterative multi-turn use cases.
- Hands-on experience in post-training multimodal models for multimodal reasoning.
Ways to stand out from the crowd:
- Strong record of publications in top-tier AI and machine learning venues such as NeurIPS, ICML, ICLR, CVPR, LLMSEC, OMMM, or WAITI
- Community Contributions: Active contributions to open-source AI safety tools, benchmarks, datasets, or models.
- Multilingual Security: Experience with multilingual AI security.
- Experience with auto red-teaming.
With highly competitive salaries and a comprehensive benefits package, NVIDIA is considered one of the industry’s most desirable employers. As you plan your future, see what we can offer you and your family at www.nvidiabenefits.com/ . If you are a creative and autonomous engineer with a genuine passion for state-of-the-art technology, we want to hear from you!
Your base salary will be determined based on your location, experience, and other objective job-related factors. The expected base salary ranges for this position are: 187,500 USD – 364,000 USD for Level 4; and 232,500 USD – 442,000 USD for Level 5.
You will also be eligible for equity and benefits .
Applications for this job will be accepted at least until November 23, 2026.
This posting is for an existing vacancy.
NVIDIA uses AI tools in its recruiting processes.
NVIDIA is committed to fostering an inclusive work environment and proud to be an equal opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.
