Lakera
Senior Research Engineer - Security Foundation Models(m/w/x)
Full-timeOn-siteSenior
Zürich
New Job?Nejo!
The AI Job Search Engine
Post-training and instruction-tuning state-of-the-art LLMs, improving Gemini Models' safety and adversarial robustness at an AI research organization. PhD in Computer Science or equivalent, with significant LLM post-training experience required. Work on cutting-edge Gemini Models, focusing on AI safety and ethics.
Requirements
- PhD in Computer Science, related field, or equivalent practical experience
- Significant LLM post-training experience
- Advantageous: Experience in Reward modeling and Reinforcement Learning for LLMs Instruction tuning
- Advantageous: Experience with Long-range Reinforcement learning
- Advantageous: Experience in Safety, Fairness, and Alignment
- Advantageous: Track record of publications (NeurIPS, ICLR, ICML)
- Advantageous: Experience taking research from concept to product
- Advantageous: Experience collaborating or leading applied research projects
- Advantageous: Strong experimental taste and good judgment
- Advantageous: Experience with JAX
Tasks
- Post-train and instruction-tune state-of-the-art LLMs
- Develop LLMs for diverse modalities and agentic capabilities
- Explore data, reasoning, and algorithmic solutions
- Ensure Gemini Models are safe, helpful, and inclusive
- Improve Gemini's adversarial robustness against high-stakes abuse risks
- Design and maintain high-quality evaluation protocols
- Assess model behavior gaps for safety and fairness
- Assess model headroom for safety and fairness
- Develop and execute experimental plans
- Address known model gaps through experimentation
- Construct new model capabilities through experimentation
- Drive innovation in Supervised Fine Tuning
- Drive innovation in Reinforcement Learning fine-tuning
- Enhance understanding of Supervised Fine Tuning at scale
- Enhance understanding of Reinforcement Learning fine-tuning at scale
Work Experience
- approx. 4 - 6 years
Education
- Doctoral / PhD
Languages
- English – Business Fluent
Tools & Technologies
- LLM
- Reinforcement Learning
- JAX
Find the original job posting in its most current version here. Nejo automatically captured this job from the website of DeepMind and processed the information on Nejo with the help of AI for you. Despite careful analysis, some information may be incomplete or inaccurate. Please always verify all details in the original posting! Content and copyrights of the original posting belong to the advertising company.
Not a perfect match?
- Intrinsic
Research Scientist, Deep Learning(m/w/x)
Full-timeOn-siteExperiencedZürich - Intrinsic
Senior AI Research Scientist - Vision-guided robotics(m/w/x)
Full-timeOn-siteSeniorMünchen, Zürich - Google DeepMind
Research Engineer, Multimodal Reinforcement Learning(m/w/x)
Full-timeOn-siteExperiencedZürich - NVIDIA Switzerland AG
Research Scientist, ML Systems - PhD New College Grad(m/w/x)
Full-timeOn-siteExperiencedZürich
Post-training and instruction-tuning state-of-the-art LLMs, improving Gemini Models' safety and adversarial robustness at an AI research organization. PhD in Computer Science or equivalent, with significant LLM post-training experience required. Work on cutting-edge Gemini Models, focusing on AI safety and ethics.
Requirements
- PhD in Computer Science, related field, or equivalent practical experience
- Significant LLM post-training experience
- Advantageous: Experience in Reward modeling and Reinforcement Learning for LLMs Instruction tuning
- Advantageous: Experience with Long-range Reinforcement learning
- Advantageous: Experience in Safety, Fairness, and Alignment
- Advantageous: Track record of publications (NeurIPS, ICLR, ICML)
- Advantageous: Experience taking research from concept to product
- Advantageous: Experience collaborating or leading applied research projects
- Advantageous: Strong experimental taste and good judgment
- Advantageous: Experience with JAX
Tasks
- Post-train and instruction-tune state-of-the-art LLMs
- Develop LLMs for diverse modalities and agentic capabilities
- Explore data, reasoning, and algorithmic solutions
- Ensure Gemini Models are safe, helpful, and inclusive
- Improve Gemini's adversarial robustness against high-stakes abuse risks
- Design and maintain high-quality evaluation protocols
- Assess model behavior gaps for safety and fairness
- Assess model headroom for safety and fairness
- Develop and execute experimental plans
- Address known model gaps through experimentation
- Construct new model capabilities through experimentation
- Drive innovation in Supervised Fine Tuning
- Drive innovation in Reinforcement Learning fine-tuning
- Enhance understanding of Supervised Fine Tuning at scale
- Enhance understanding of Reinforcement Learning fine-tuning at scale
Work Experience
- approx. 4 - 6 years
Education
- Doctoral / PhD
Languages
- English – Business Fluent
Tools & Technologies
- LLM
- Reinforcement Learning
- JAX
Find the original job posting in its most current version here. Nejo automatically captured this job from the website of DeepMind and processed the information on Nejo with the help of AI for you. Despite careful analysis, some information may be incomplete or inaccurate. Please always verify all details in the original posting! Content and copyrights of the original posting belong to the advertising company.
About the Company
DeepMind
Industry
Science
Description
The company advances the state of the art in artificial intelligence for public benefit and scientific discovery.
Not a perfect match?
- Lakera
Senior Research Engineer - Security Foundation Models(m/w/x)
Full-timeOn-siteSeniorZürich - Intrinsic
Research Scientist, Deep Learning(m/w/x)
Full-timeOn-siteExperiencedZürich - Intrinsic
Senior AI Research Scientist - Vision-guided robotics(m/w/x)
Full-timeOn-siteSeniorMünchen, Zürich - Google DeepMind
Research Engineer, Multimodal Reinforcement Learning(m/w/x)
Full-timeOn-siteExperiencedZürich - NVIDIA Switzerland AG
Research Scientist, ML Systems - PhD New College Grad(m/w/x)
Full-timeOn-siteExperiencedZürich