Lead Machine Learning Engineer, Multimodality, Gemini

Google DeepMind is seeking a Lead Machine Learning Engineer to spearhead the development and deployment of cutting-edge multimodal AI models for Gemini in Zurich, Switzerland. This role will directly shape the future of how users interact with image, video, and audio content across Google's platforms. The ideal candidate will possess a deep understanding of SOTA research and the ability to translate it into scalable, production-ready solutions.

Responsibilities:

• Collaborate with research teams to evaluate and drive SOTA multimodal technologies.
• Provide technical leadership in defining the strategic direction for multimodal model development.
• Lead the implementation of advanced techniques, including SFT, RL*F, IPO/DPO, to drive quality improvements.
• Independently design and execute complex experiments to validate and refine model architectures.
• Conduct rigorous data analysis to identify insights and opportunities for enhancing multimodal capabilities.
• Develop data-driven recommendations to inform the strategic development of a robust data flywheel.
• Act as a technical mentor to other team members.

Requirements:

• Master's degree or PhD in Computer Science, AI, Machine Learning, or a related field.
• Proven experience in developing and deploying large-scale machine learning models, particularly in multimodal AI.
• Experience with large language models and multimodal model architectures.
• Strong problem-solving and analytical skills, with the ability to design and execute complex experiments.

The role offers:

• Opportunity to work on cutting-edge multimodal AI development for Gemini.
• Chance to contribute to a strategic priority for Google, aligned with industry-wide trends.
• Exposure to advanced techniques like fine-tuning, RL*F, and PolicyOptimization / PreferenceData.