Lead Scientist - Post-Training AI Agents for Consumers

Google LLC; Mountain View, Kirkland (CA, WA); Hybrid; USD 262,000 - 364,000; Full time

Benefits offered by this job

Health insurance

Dental insurance

Vision insurance

Life insurance

Disability insurance

401(k) with company match

Paid time off

Sick time

Maternity leave

Baby bonding leave

Holidays

Job summary

Google DeepMind in Mountain View, CA seeks a Senior Staff Research Scientist/Engineer focused on Consumer Agents Post Training. You will research, develop and scale post-training methods including large language models, RL and personalization to build autonomous consumer agents.

The role emphasizes modeling and research leadership, with collaboration across teams, and opportunities to contribute to scalable evaluation and datasets for long-horizon reasoning.

Qualifications

• Bachelor’s degree or equivalent practical experience.
• 8 years of software development experience.
• 5 years of design and architecture experience with testing/launching software.
• Experience with large language models, reinforcement learning and deep learning.

Responsibilities

• Drive core modeling and post-training research for agent post-training.
• Design and execute ablation studies to validate hypotheses.
• Communicate results through visualizations and reports.
• Develop research infrastructure and data/model evaluation tools.
• Collaborate with researchers and engineers to maintain feedback loops.

Skills

LLM

Reinforcement Learning

Deep Learning

Education

Bachelor’s degree or equivalent

Master’s degree or PhD

Job description

Google DeepMind in Mountain View, CA seeks a Senior Staff Research Scientist/Engineer focused on Consumer Agents Post Training. You will research, develop and scale post-training methods including large language models, RL and personalization to build autonomous consumer agents.

The role emphasizes modeling and research leadership, with collaboration across teams, and opportunities to contribute to scalable evaluation and datasets for long-horizon reasoning.