Research Engineer, Gemini Retrieval, DeepMind

Minimum qualifications

• Bachelor's degree in Computer Science, Data Science, Artificial Intelligence, Machine Learning, or a related technical field, or equivalent practical experience.
• 3 years of experience in machine learning, with a focus on Large Language Models (LLMs) or Information Retrieval (IR).
• 3 years of experience with software engineering in Python, C++, JAX, or PyTorch programming, including experience implementing multi-stage training pipelines or algorithms.

Preferred qualifications

• Experience developing Reinforcement Learning (RL) algorithms, automated evaluation systems, or LLM reflection and reasoning frameworks.
• Experience with competitive programming, mathematics competitions (e.g., ICPC, IOI, IMO), or a track record of developing novel algorithms.
• Experience rapidly prototyping and iterating on complex systems, generative AI models, or multi-stage pipelines.
• Experience collaborating with product engineering teams to deploy machine learning models or research innovations into large-scale production environments (e.g., Search, Recommendation Systems).
• Track record of learning new software tools, frameworks, and codebases quickly to solve complex technical obstacles.

About the job

Research-focused Software Engineers create experiments, prototype implementations, and design new architectures for real-world challenges such as artificial intelligence, data mining, natural language processing, hardware and software performance analysis, compilers for mobile platforms, and core search. The role includes large-scale testing and rapid deployment of promising ideas, while contributing to the wider research community through university partnerships and published papers.

Responsibilities

• Uncover strategic opportunities at the intersection of Gemini, Search, factuality, continual learning, deep research, and domain internalization.
• Analyze models and model-driven products beyond current leaderboards; understand complex problem spaces and create new ways to measure ideal behavior.
• Use empirical findings to develop practical interventions and modeling innovations that improve models across downstream surfaces.
• Collaborate with product teams such as Search and YouTube to scale research innovations into production environments, optimizing for quality and efficiency.