Sr Manager Data Science - GenAI Labeling Platforms

Pinterest brings millions of people the inspiration to create a life they love. To evolve this mission our product and engineering teams need to innovate much of this innovation is driven through machine learning and critical to understanding how innovation on the product will change our business is through measurement. Data labels play a key role in unlocking both of these and the rate and scale at which we can obtain reliable labels limits our speed of innovation.

This role brings together 3 critical science efforts:

• The notion of asking qualified raters to answer questions about content whether its scoring its relevance to a particular context review captions for pins or determining whether or where certain attributes are present in an image has played a critical role in our evaluation and machine learning efforts across Pinterest. However utilizing human evaluators to complete these tasks at scale have limitations it can be timeintensive to accurately define the tasks costly to get large volumes of labels and slow to collect the data.
• Surveys of our Pinners offer a unique mechanism to gather the perspective of our Pinners on questions we have about the platform and content; a Pinners perspective can be different to that of a human rater observing the same content because of the unique context they possess.
• Advancements in Generative AI have opened up a wealth of opportunities for improvements in productivity and weve only scratched the surface of its capabilities in the labeling space. Early results show strong promise for the role it can play in these tasks reducing the time and cost of our labeling efforts focusing our surveys and human rater efforts on higher value problems and improving the accuracy of our learnings.

Were looking for an accomplished Data Science leader to lead a small highperforming team to unlock a new level of capabilities in our efforts across labeling surveys and iterative GenAI tasks. From statistical approaches that can better inform the labels we collect to prototypes to automatically craft prompts for LLMs to address the needed tasks this team will lead a key strategic component of the data science organization innovating on data platforms across the company.

What youll do:

We are looking for an experienced and highly capable Data Scientist to help us drive step function improvements in our data labeling capabilities at Pinterest. In this role you will:

• Create and realize a strategic vision for the science investments across the labeling survey and GenAI tooling space emphasizing rigor alongside valuecreation and partner with product and engineering leaders to build a unified vision for these platforms.
• Build inspire and grow a small high performing team of data scientists who are passionate about this space that will shape and execute on the strategic vision.
• Apply deep scientific knowledge and domain expertise to understand opportunities and prototype solutions that demonstrate substantial improvements in velocity accuracy and costefficiency measure and mitigate biases and amplify the impact employees can have when utilizing survey results data labels or GenAI prompts in their investments.
• Translate ambiguous business challenges into actionable plans and ensure that your team is consistently producing trustworthy and highquality technical output that influence the business.
• Communicate complex analytical findings and insights to both technical and nontechnical audiences in a clear and concise manner.
• Collaborate crossfunctionally with customers to understand their problems and painpoints and with engineers to scale successful prototypes to become integral components of the platform.

What were looking for:

• 10 years of combined postgraduate academic and industry experience applying scientific methods to solve realworld problems on largescale data
• Experience in managing teams of data scientists fostering a culture of innovation high performance collaboration and continuous improvement
• Handson experience as an individual contributor using a deep understanding of scientific methods applied to data to drive business decisions
• Demonstrated and impact on initiatives that cross multiple product areas and interface with leadership and product teams with a track record of influencing leaders and peers using data
• Proven track record of crafting high quality algorithmic code and identifying opportunities for efficiency and performance improvements through statistical methods
• Demonstrated and impact on crossfunctional initiatives strong communication skills and a track record of influencing leaders and peers using data
• Selfpropelled continuous learner who keeps up with new tools and methodologies and builds prototypes with concepts learned
• Strong business and product sense with the ability to shape ambiguous questions into welldefined analyses and success metrics that drive business decisions
• The ideal candidates will have experience with labeling platforms measurement and developing high quality prompts for LLMs

Relocation Statement:

• This position is not eligible for relocation assistance. Visit our PinFlex page to learn more about our working model.
• This role will need to be in the office for inperson collaboration 35 times/quarter and therefore can be situated anywhere in the country.

#LINM4

#LIREMOTE

Required Experience:

Manager