Job Responsibilities
As a Staff / Principal Research Scientist, you will:
- Frame research questions and design experiments to find answers.
- Take full-stack ownership of research projects, from conception to publication.
- Evaluate and measure the quality of models and experiments.
- Collaborate with a team to tackle unclear problems and develop solutions.
- Contribute to the development of multimodal models and agentic systems.
- Support the sharing of work that advances the field of AI.
Requirements
- PhD in Machine Learning or Natural Language Processing, or equivalent practical experience.
- Experience with foundation models, evaluation methods, and frontier topics in AI.
- Published research in reputable conferences such as ICML, ICLR, NeurIPS, EMNLP, ACL, or AAAI.
- Demonstrated ability to conduct non-trivial AI side projects or open-source contributions.
Qualifications
- Strong problem-solving skills and ability to work in ambiguous situations.
- Experience in designing experiments and evaluating results.
- Ability to communicate complex ideas clearly and effectively.
Benefits
- Base salary range of $270,000 - $500,000+ bonus + equity.
- Relocation assistance for candidates moving to Mountain View.
- Support for sharing research work that contributes to the field.
Perks
- In-person collaboration to foster a strong team culture.
- Flat organizational structure with fast iterations and minimal process.
What You'll Do
You will engage in cutting-edge research, develop innovative AI models, and contribute to the advancement of the field through impactful work.
Who You Are
You are a proactive researcher who thrives in ambiguity, values impact over academic output, and is driven by a desire to understand the fundamental logic behind decisions.
Tech Stack
Experience with machine learning frameworks and tools relevant to foundation models and multimodal systems is preferred.
Team Description
You will be part of a product-oriented research lab comprised of top AI researchers and engineers, focused on developing best-in-class real-time multimodal models.
