Surabhi Seth Email & Phone Number

United States

Greater Seattle Area

Saved thousands of dollars in marketing cost by creating a deep learning model to enable optimal marketing spend allocation across paid search and social that lowered Cost per Qualified Lead (CPQL) for paid search by 18%.Discovered key early indicators of customer churn and developed a highly accurate tree-based churn prediction model to identify customers at risk.Identified limitations in the existing ARR calculation method and developed a robust solution less prone to volatility than legacy methods, that enhances business transparency and simplifies communication for board members.As the first data science hire at Boulevard, architected a playbook for model development, hypertuning and deployment using cutting edge tools like Snowpark and dbt.

Kirkland, Washington, United States

Built and deployed a bid optimizer to optimize spend on search engine marketing (SEM) using deep neural networks, integer programming and Airflow. As the team lead, steered project planning, mentored team members, and ensured smooth cross-functional collaboration with the marketing team.Developed a classification model utilizing domain names to predict region of operation, enabling data-driven quantification of regional domain market share growth for C-level reporting.

Phoenix, Arizona Area

Created machine learning models to predict which customers were most likely to choose a competitor for their next venture.For marketing suitable product offerings and measuring the impact of such marketing campaigns, identified unique segments (using clustering) from customers at risk of migrating to a competitor.

Gurgaon, India

I worked with many mid to large size clients on a variety of domains like advertising, manufacturing and retail. My interest in analytics grew while developing a business intelligence product for a Fortune 500 client which contributed in plugging leakages in the aftermarket sales funnel by providing predictive maintenance leads. There I also designed a method which was able to classify 90% of the unaccounted parts sales and which would contribute an immense improvement in C-level reporting. Overall, I carried following responsibilities during the course of different engagements:• Ran discovery workshops to understand clients’ ecosystem and pain points• Analyzed ‘as-is’ states and laid down the ‘to-be’ user journeys• Analyzed enterprise data to validate business use cases• Created product backlogs and worked with clients and the development teams to synthesize release plans• Showcased product features to client subject matter experts and conducted focus groups for ideation and feedback• Led and mentored agile teams to deliver faster and better

Master Of Science - Ms, Machine Learning And Data Science

Bachelor Of Technology - Btech, Computer Science And Engineering