Research Scientist, Reasoning and Artificial General Intelligence, DeepMind

Google creates products and services that make the world a better place, and gTech’s role is to help bring them to life. Everything we do aims to ensure our customers benefit from the full potential of Google products.

gTech Risk helps protect Google and its users by identifying and implementing controls for significant risks. To ensure gTech's security and agility, the team analyzes gTech business processes to find control gaps, collaborates with stakeholders to resolve these gaps, and provides risk expertise to engineer solutions.

gTech Risk Solutions Engineering is executing a strategic transformation from a reactive manual process led organization to a proactive engineering powerhouse. Operating as a federated security partner to Google's central security organizations, we design and build contextualized security solutions tailored specifically for the gTech business ecosystem. Our mission is upfront risk prevention through technical delivery excellence, shifting the paradigm from unsustainable "risk-chasing" to "prevention by design."

In this role, you are the observability and feedback engine of the team. You will ensure our autonomous remediations and agentic workflows are operating securely and effectively.

You will design and monitor the automated flows of agentic tooling and manage the operational fallback tiers when AI workflows flag anomalies or cannot manage required strategies for gTech's Access ecosystem. Utilizing your deep security domain knowledge, you will conduct actor and account access analysis, develop strategies for least privliedge, perform complex log analysis (SQL), and triage incident alerts. Crucially, your ground-level intelligence will feed directly back into our engineering roadmap, converting identified gaps and opportunities into systemic, automated prevention controls. Individual pay is determined by factors including job-related skills, experience, and relevant education or training.

US: $123000 - $176000 (USD) + 15% bonus target + equity + benefits

Learn more about benefits at Google. Responsibilities

• Conduct critical research in RL and reasoning by proposing new RL algorithms, scaling RL or inference time compute for improving Gemini.

• Serve as Gemini launch captains, acting as launch/model captains and training models to directly contribute to Gemini landings.

• Anticipate where the next game changers will come from, staying mindful of upcoming research breakthroughs, and build research agendas based on new paradigms to execute on them.

• Execute on promising research agendas by working as a team.

• Collaborate with other Gemini core teams, mostly based in the US and UK.

Qualifications Minimum qualifications:

• PhD in Computer Science, a related field, or equivalent practical experience.

• 4 years of experience with research agendas across multiple teams or projects.

• Experience conducting research in Reinforcement Learning (RL) or reasoning.

• Experience with Large Language Models (LLM).

• One or more scientific publication submission(s) for conferences, journals, or public repositories.

Preferred qualifications:

• 2 years of experience in coding and leading multiple research efforts and influencing research direction.

• Experience driving research projects from conception to completion.

• Ability to work in a fast-paced, dynamic, and collaborative environment.

• Excellent problem-solving, analytical, and communication skills.

• A track record of research or engineering achievements, including publications in peer-reviewed conferences or journals.