Research Engineer in Gemini Post Training

Google DeepMind is seeking a highly motivated and talented Research Engineer to join their team at the forefront of post training research in Gemini. This role involves working alongside a world-class team of researchers and engineers to develop and advance the next generation of frontier AI models.

Key responsibilities include:

1. Post-training for instruction following and tool use: Develop and apply methods to enhance the instruction following abilities of Gemini models targeting next-generation capabilities. This may include fine-tuning strategies, reinforcement learning techniques, or novel approaches to improve Gemini's ability to solve complex tasks, make decisions, and interact with dynamic environments.
2. Develop and evaluate next-generation applications of GenAI: Translate research findings into practical applications by building and evaluating prototypes that stretch instruction following model behaviors that affect user-facing scenarios. Leverage learnings from such use cases and transform them into post training methods that can improve the Gemini model.

The ideal candidate should have:

• A degree in Computer Science, Statistics, or a related field
• Expertise in one or more areas: machine learning, natural language processing, computer vision, reinforcement learning
• Experience with training, evaluating, or interpreting large language models (advantage)
• Proven ability to design and execute independent research projects (advantage)
• Excellent communication and collaboration skills (advantage)
• Extensive experience with deep learning frameworks and large-scale model training (advantage)

Google DeepMind values diversity and is committed to equal employment opportunity. They offer a competitive salary range and benefits package. The application deadline is November 15th, 2024.