Frontier AI Research Engineer in Montréal

Google DeepMind is one of the world's leading AI research labs, and this Research Developer role sits right at the frontier of what's possible. Based in Montreal, QC, you'll be embedded within the Frontier AI Incubation team — a group pushing the boundaries of foundational AI systems across multiple domains, from personalized model adaptation to multimodal reasoning and agentic workflows.

This isn't a role where you're maintaining existing pipelines. You'll be doing end-to-end technical development — designing algorithms, running experiments, building production-grade architectures, and partnering directly with Google product and development teams to bring cutting-edge AI capabilities into real-world environments. The work spans research and product, and you'll be expected to move fluidly between both.

At the core of this position is the design, training, and optimization of foundational algorithms and machine learning systems. That includes working on personalized model adaptation, contextual memory architectures, dynamic prompt optimization, and multimodal reasoning. You'll lead technical development from early algorithmic design through to production-grade architecture and scaled serving infrastructure, ensuring that promising research ideas can be deployed broadly and quickly.

Collaboration is central to how this team operates. You'll partner directly with Google development and product teams to integrate core technologies into production environments — including Project Helix, agent workspaces, and intelligent system integrations. You'll also be responsible for formulating novel evaluation methodologies, both automated and human-in-the-loop, to measure capability gains, latency and compute efficiency, alignment, and personalization fidelity.

Responsibilities:

• Design, train, and optimize foundational algorithms and ML systems, including personalized model adaptation, agentic workflows, contextual memory architectures, dynamic prompt optimization, and multimodal reasoning.
• Lead end-to-end technical development from algorithmic design and experimental prototyping through to production-grade architecture and scaled serving infrastructure.
• Partner directly with Google development and product teams to integrate and harden core technologies within production environments, including Project Helix, agent workspaces, and intelligent system integrations.
• Formulate novel evaluation methodologies, both automated and human-in-the-loop, to measure capability gains, latency and compute efficiency, alignment, and personalization fidelity.

Requirements / Skills:

• Bachelor's degree in Computer Science, Machine Learning, Mathematics, Statistics, or equivalent practical experience.
• 2 years of experience in machine learning, algorithm design, data structures, or distributed software systems.
• 2 years of programming experience in Python or C++.
• 1 year of experience taking technical projects or ML systems from conceptual formulation through to implementation and deployment.
• Experience with foundation model development, including techniques such as reinforcement learning from human/AI feedback, supervised fine-tuning, parameter-efficient tuning, or inference optimization (preferred).
• Experience or interest in personalization, adaptive systems, user modelling, retrieval-augmented generation, or agentic memory architectures (preferred).
• Experience with ML frameworks and large-scale model training or serving infrastructure (preferred).
• Experience collaborating across research and product boundaries to co-design technical architectures (preferred).
• Strong scientific analysis, technical problem-solving, and communication skills (preferred).