Research Engineer, Agentic AI at DeepMind

Mountain View, California, United States -

Full Time

Start Date

Immediate

Expiry Date

05 Jan, 26

Salary

0.0

Posted On

07 Oct, 25

Experience

2 year(s) or above

Remote Job

Yes

Telecommute

Yes

Sponsor Visa

Skills

Machine Learning, Python, Deep Learning, Reinforcement Learning, Natural Language Processing, Agent-Based Systems, Large Language Models, C++, Model Training, Data Orchestration, Benchmarking, AI Product Development, Long-Context Reasoning, Multi-Agent Systems, Tool Use, Algorithm Design

Industry

Research Services

Description

Snapshot We are the team that built and shipped Google Search's AI Mode from scratch to production. Now, we are embarking on our next mission to build the future of AI assistants. As a Research Scientist or Engineer on this team, you will be at the forefront of AI, tackling foundational challenges in agentic reasoning, planning, and tool use, with a clear and proven path to delivering your work to billions of users. About Us At Google DeepMind, we’re a team of scientists, engineers, machine learning experts and more, working to advance the state of the art in artificial intelligence and ship AI products at a global scale. From data curation, model training, evaluation to serving orchestration, we own end to end development of a helpful AI product. Our team has a proven track record of applying innovative techniques and creating multiple AI products from inception. The Role Your work will directly address some of the most critical open questions in AI today. How do we build agents that can manage long-context dialogues and tasks? How can an agent learn to orchestrate diverse tools—from web search and browsing to UI interaction and private data access—efficiently and effectively? How do we scale from a single powerful agent to a collaborative multi-agent system? You will be responsible for developing novel solutions to these questions, implementing them within our advanced research stack, and benchmarking their performance on challenging, real-world tasks. This is a unique opportunity to build the core components of an AI that can help a user act as an expert shopping companion, help design a living room or planting planner through conversation. You will contribute to a common technical architecture that connects cutting-edge capabilities like bidirectional streaming and long-context, private data retrieval, and Deep Research, with a clear path to bring successful experiments to millions of users. Key responsibilities: Design, implement, and iterate on novel models and algorithms (e.g., Agentic RL) to improve agentic tool use. Tackle the core research problem on long-context reasoning. Develop sophisticated agentic orchestration from data, model training to inference. Explore and build advanced multi-agent systems in both Python for rapid experimentation and C++ for high-performance deployment. Continuously evaluate and benchmark the models against industry standards. Partner with product and engineering teams to integrate your research breakthroughs into live experimental product surfaces, directly influencing the future of Google's assistants. About You In order to set you up for success as a Research Scientist / Research Engineer at Google DeepMind, we look for the following skills and experience: Requirements: Bachelor's or Master's degree in Computer Science, a related technical field, or equivalent practical experience. 3+ years of professional or academic experience in machine learning or a related field. Strong programming skills and experience in Python. Experience with one or more deep learning frameworks (e.g., JAX, TensorFlow, PyTorch). Experience in one of the following research areas: Reinforcement Learning, Large Language Models (LLMs), Agent-based Systems, or Natural Language Processing. In addition, the following would be an advantage: A PhD in Computer Science, Machine Learning, or a related field, with a strong publication record in top-tier conferences. Demonstrated experience building and training agentic AI systems, particularly with tool use (e.g., ReAct-style models). Deep expertise in Reinforcement Learning, especially in the context of large-scale models and long-horizon tasks. Experience with long-context models, context compression techniques, or state-space models. Proficiency in Python/C++ for building high-performance, low-latency machine learning systems.

Responsibilities

You will develop novel solutions to foundational challenges in AI, focusing on agentic reasoning, planning, and tool use. Your work will involve designing and implementing models and algorithms, as well as collaborating with product and engineering teams to integrate research breakthroughs into live products.