As a Principal Data Scientist, you will be a deeply technical and strategic data science leader for Gemini Core Modeling and Evaluations. You (along with a small team of data scientists) will be responsible for defining the model evaluation criteria, influencing the supervised fine tuning data and evaluation prompt sets, identifying valuable signals for reinforcement learning human feedback, and improving human rater efficiencies. The US base salary range for this full-time position is $262,000-... more details
As a Principal Data Scientist, you will be a deeply technical and strategic data science leader for Gemini Core Modeling and Evaluations. You (along with a small team of data scientists) will be responsible for defining the model evaluation criteria, influencing the supervised fine tuning data and evaluation prompt sets, identifying valuable signals for reinforcement learning human feedback, and improving human rater efficiencies.
The US base salary range for this full-time position is $262,000-$377,000 + bonus + equity + benefits. Our salary ranges are determined by role, level, and location. The range displayed on each job posting reflects the minimum and maximum target salaries for the position across all US locations. Within the range, individual pay is determined by work location and additional factors, including job-related skills, experience, and relevant education or training. Your recruiter can share more about the specific salary range for your preferred location during the hiring process.
Please note that the compensation details listed in US role postings reflect the base salary only, and do not include bonus, equity, or benefits. Learn more about benefits at Google.
Develop deep product and user analytics that will help to drive quality and design directions, such as identifying signals that are predictive of user engagement or analyzing Gemini use cases (e.g., coding, fact finding, or creativity and summarization). Develop insights around trends and drive model improvements that help achieve goals.
Drive forward the science of measuring Gemini, ensuring that our quality, factuality, and safety model evaluation metrics are indicative of the end-user experience, and help guide launch decisions for Gemini.
Manage experimentation and comparative evaluation, including the science, methodologies, tools, and measurement approaches to accelerate Gemini experimentation and innovation.
Guide the model's human evaluation process, including analyzing rater performance, influencing rater templates, updating training, and evolving our approach to ensure precise and accurate evaluations.
Minimum qualifications:
Bachelor's degree in Machine Learning, Data Science, Information Management, Computer Science, or other Engineering discipline, or equivalent practical experience.
15 years of experience in data science or other quantitative fields such as: statistics, machine learning, software engineering, or ML engineering.
Experience working on algorithmic products, production machine learning systems, and defining model evaluation metrics.
Experience in using data insights to drive product excellence.
Preferred qualifications:
Experience developing high-growth technology products and services.
Ability to influence without direct authority, working on and with technical and cross-functional or highly matrixed teams who do not report into this role.
Ability to turn ambiguous problem spaces into clear solutions, and comfortable with new technology and thinking outside the box to develop and implement short- and long-term creative solutions.
Ability to work well cross-functionally, and manage organizational complexities and competing priorities.
Excellent project/product management and technical architecture skills, and the ability to deliver projects on time.
Excellent strategic thinking, data science, and analytical skills.
Job Abstracts is an independent Job Search Engine. Job Abstracts is not an agent or representative and is not endorsed, sponsored or affiliated with any employer. Job Abstracts uses proprietary technology to keep the availability and accuracy of its job listings and their details. All trademarks, service marks, logos, domain names, and job descriptions are the property of their respective holder. Job Abstracts does not have its members apply for a job on the jobabstracts.com website. Additionally, Job Abstracts may provide a list of third-party job listings that may not be affiliated with any employer. Please make sure you understand and agree to the website's Terms & Conditions and Privacy Policies you are applying on as they may differ from ours and are not in our control.
We would like to take a second to Welcome You to Job Abstracts, the nation’s largest Pure Job Board. With over 3.1 million job listings from 15,000+ Companies & Organizations, we help job searchers find careers that match their interests. As an anonymous user, you have probably discovered how easy our system is to use. However, you have just scratched the surface of what we can offer.
We encourage you to Register so you can use our most powerful features: searching with multiple terms, setting up multiple locations, establishing favorite companies, and accessing your search history. If you find a job you like, you can apply directly for it, and then, keep notes on it. We will also keep a lookout for jobs that match your search terms and email you when we find something you may like.
You can register for free and the system is free to use. If you like our system so far, click on Register and unlock the power required by serious job searchers.
Any time you conduct a search, the system shows you job matches, ranked by their Relevance Score (RS).
The score is calculated by a proprietary algorithm that uses Intelligent Machine Learning.
The Relevance Score tells you how well the job opportunity matches your search term or terms.
When not logged in, the system is limited to one search term. Scores for single term matches are usually lower.
When you register, log in, and set up multiple terms prioritized by importance, the jobs found for you will receive a much higher Relevance Score.