At Handshake, I built out the Data / AI org and identified the highest leverage opportunities. We have focused on improving job search and recommendations, automating audience building of our flagship monetization product, Campaigns, a smart notifications platform to drive engagement thoughtfully, and retrieval and ranking services to power the user feed. More recently, our focus has been on leveraging Generative AI technologies to power core experiences for both employers and early career talent on the Handshake platform.

I led Artificial Intelligence initiatives for multiple business units (ads, jobs, and learning). I was fortunate to lead a team of outstanding scientists and engineers who developed personalized data products and contributed to LinkedIn's IP portfolio through research. We work on fundamental AI technologies across these marketplaces (e.g., hyper-personalization of courses, jobs, ads, etc., trade-offs among revenue, engagement, and other objectives, algorithmic bidding, incremental learning on data-streams, large scale allocation of jobs/ads to members) My team and I also were instrumental in powering the launch of new LinkedIn products (e.g., LinkedIn Learning: Re-skilling the global workforce, P4P jobs: transforming our subscription-based jobs business to a pay-for-performance model). Open-source code:DuaLip: [https://github.com/linkedin/DuaLip/](https://github.com/linkedin/DuaLip/)Lambda Learner: [https://github.com/linkedin/lambda-learner](https://github.com/linkedin/lambda-learner)Recent Research papers:LAWN: [https://arxiv.org/abs/2108.05839](https://arxiv.org/abs/2108.05839)Marketplace: [https://arxiv.org/abs/2103.05277](https://arxiv.org/abs/2103.05277)Incremental Learning: [https://arxiv.org/abs/2010.05154](https://arxiv.org/abs/2010.05154) , Video:  [https://youtu.be/zSlZwQ4Tf4w](https://youtu.be/zSlZwQ4Tf4w)Deep learning for search: [https://arxiv.org/abs/1809.06473](https://arxiv.org/abs/1809.06473)Engineering blog:https://bit.ly/3mz83pn https://bit.ly/3BdyBk5 https://bit.ly/3kqSRbdhttps://bit.ly/3zpNuzp

I developed Machine Learning and Statistical tools in search advertising effectiveness and to help improve the targeting of Google's ads products and sales offerings.I attempted to answer the following questions using experimental and observational data: (1) How can sales teams optimize their portfolio of customers? (2) How should sales and marketing teams prioritize among a large portfolio of advertisers in their book of business? I worked with a cross-functional team comprising product, eng, sales, and marketing in formulating the business problems, building and deploying the statistical models, and for broader communication around the impact and rollout. My work on using predictive models to help the sales organization prioritize the portfolio of advertisers was recommended for immediate deployment by exec leadership and was featured in the letter to the board of directors. Later in my tenure at Google, I led research in the area of Search Ads Effectiveness. The goal for my team was to develop and deploy @ scale, research solutions for measuring search advertising. -- Search Ads Pause: Observational Ad Effectiveness Platform to estimate impact of paid search on site traffic, and to quantify the effect of organic rank on paid search incrementality. This pipeline produces tens of thousands of studies and monitors a substantial portion of Google's search ads revenue. It has helped us publish ground-breaking industry benchmarks. -- Geo-experiments Platform: Experimental ad-effectiveness platform to answer various attribution questions on the effectiveness of search ads. Examples: determining optimal keyword coverage on brand/generic terms, finding efficient bidding strategies, quantifying the impact of online ads on offline store-sales.

* Developed market forecasting models for applications in product design using various advanced statistical modeling techniques (e.g., Mixed Logit, Nested Logit, Hierarchical Bayes, Probit). * Collaborated with Ford Motor Co. and JD Power Associates on the development of market models for the automobile market. * Developed efficient computational methods to support the research effort.