Senior ML Research Scientist

The role

You will design and implement innovative ML models to solve advanced AI safety problems. You will collaborate with mathematicians and other researchers to integrate theoretical advancements into practical algorithms.

What you will do

• Propose, design, and implement novel ML models tailored to solve complex AI safety problems.
• Collaborate with mathematicians and other specialized research scientists to integrate theoretical advancements into practical ML algorithms and models.
• Adapt and fine-tune existing frontier models to enhance their applicability to specific settings.
• Design and implement experimental protocols and evaluation frameworks to validate hypotheses.
• Analyze and interpret experimental data to steer future research questions and objectives.
• Benchmark and optimize model performance and efficiency with ML engineers.
• Communicate findings and synchronize work with other researchers and engineers.

Required qualifications

• Advanced degree in a relevant field (e.g., Computer Science, Mathematics).
• 5+ years of experience in deep learning research projects, in particular with frontier models.
• Proven experience in the training, adaptation, and/or fine-tuning of frontier models across a variety of complex scenarios that involve, for example, transfer learning, domain adaptation, or meta lear
• Expertise in the integration and use of ML libraries such as PyTorch, TensorFlow, or JAX for the development, training, and evaluation of ML models, ideally in distributed computing environments.
• Strong experience in Python software development, in the use of version control, collaboration tools, and experiment management and tracking tools.
• Strong communication skills, both written and verbal, with the ability to exchange ideas with researchers across different disciplines.
• Ability to work collaboratively in a team environment while also being self-motivated and independent.

Nice to have

• PhD is preferred but not required if the candidate demonstrates exceptional abilities.
• Experience analyzing the behavior of frontier models in safety-critical contexts.
• Experience in the design and development of model alignment strategies or benchmarks.
• Experience with natural language processing or probabilistic graphical models.
• A track record of contributing to high-quality research projects in deep learning.

About the company

LawZero is a non-profit organization committed to advancing research and creating technical solutions that enable safe-by-design AI systems. Its scientific direction is based on new research and methods proposed by Professor Yoshua Bengio.