Senior ML Research Scientist

Position Overview

LawZero is looking for a Senior Machine Learning Research Scientist to join their team in Berlin. In this role, you will create and execute cutting-edge ML models designed to tackle sophisticated AI safety challenges.

Key Responsibilities

• Develop and implement innovative ML models specifically crafted to address intricate AI safety issues.
• Work closely with mathematicians and specialized researchers to translate theoretical insights into practical ML algorithms and techniques.
• Customize and improve existing advanced models to fit unique use cases effectively.
• Design rigorous experimental methodologies and evaluation systems to test hypotheses and guarantee reproducible, reliable outcomes, even at frontier model complexity scales.
• Interpret experimental results to guide future research directions.
• Collaborate with ML engineers to benchmark and optimize models for strong performance and efficient use of computational resources during extensive experiments.
• Effectively communicate discoveries and coordinate efforts with other researchers and engineering teams.

Qualifications and Skills

• Possession of an advanced degree in Computer Science, Mathematics, or a related discipline; a PhD is highly preferred, though exceptional candidates without one will be considered.
• At least five years experience leading deep learning research projects, especially involving advanced or frontier models.
• Demonstrated expertise in training, adapting, or fine-tuning frontier models across challenging scenarios such as transfer learning, domain adaptation, or meta learning.
• Proficiency in ML frameworks including PyTorch, TensorFlow, or JAX, particularly within distributed computing setups for model development, training, and evaluation.
• Strong foundation in Python programming and familiarity with version control systems, teamwork collaboration tools, and experiment tracking platforms.
• Excellent written and oral communication aptitude, enabling interdisciplinary knowledge exchange.
• Capable of working both independently and collaboratively within a team environment.

Preferred Qualifications

• Experience examining frontier model behaviors in environments where safety is paramount.
• Background in creating model alignment techniques or benchmarking standards.
• Skills in natural language processing and probabilistic graphical models.
• Track record of contributing to significant deep learning research projects.

About LawZero

LawZero is a non-profit that drives research and technological advances to build AI systems designed with safety at their core. Guided by Professor Yoshua Bengio, a leading AI scientist, their goal is to create AI that understands its environment rather than acts in it, offering transparent and truthful insights through probabilistic reasoning. These AI systems aim to support scientific progress, oversee autonomous AI agents, and deepen understanding of AI risks. LawZero operates out of Montreal and views AI as a global public asset to be developed responsibly and safely for humanity’s benefit.

Diversity and Privacy

LawZero values an inclusive, respectful workspace that embraces diversity. Applications from candidates committed to their mission in a collaborative and open culture are encouraged. Personal data collected during recruitment will be handled according to privacy regulations and LawZero's Privacy Policy, including rights to access and correction.