San Francisco Bay Area
I am a technical leader with experience in notifications, feed and ads ranking. I combine strong algorithmic foundations with a sharp product focus. Trusted to lead business-critical initiatives across organizations, I consistently delivered pivotal results by aligning machine learning innovation with strategic product goals.
Working on ranking problems at Instagram core ML team.
Shaped the ML strategy and execution for several high-impact company priorities: Led ranking efforts across notifications, Netflix Live, Ads and Growth, partnering with product, infra, data science and research orgs. Notable wins from these efforts: Notification arm of Netflix Live delivered multiple orders of magnitude engagement gains, and LTV modeling work that informed executive decision-making around the Ads business. Selected algorithmic details: Designed and deployed causal inference models using neural nets, improved LTV models, came up with cold-start strategies and multi-task optimization frameworks powering personalization at scale for notifications, growth and member valuation.
Worked on Pinterest homefeed and ads ranking teams. I used my expertise in deep learning to provide Pinterest users an engaging homefeed experience to help them discover and do what they love. Particularly worked on multi-task learning, model calibration, video distribution and negative signal modeling. Previously I worked as one of the founding members of the Pinterest search ads ranking team, where I have worked on candidate selection and ranking models. I have changed the model architecture of the search ads models to handle feature hash collisions in a wide model and implemented position de-biasing in the ranking model.
Working in Operational Intelligence team for Azure teams. Helped to form a new data mining and diagnostic team from ground up. Guided teams to identify their data needs, come up with metrics and instrument required logging APIs into their code base. Built batch, near real time and real time big data pipelines.
Worked as part of Bing Ads Data Mining and Diagnostics team. Mined and analyzed the data generated by the Bing Ads layer cake to define appropriate metrics and surface actionable in- formation. Investigated anomalies in data for root cause analysis. Built predictive models for live site incident prediction. Worked on backend of visualization tools and implemented analysis capabilities for engineers in these tools.
I interned at Microsoft Research Redmond as a research intern to analyze the effects of distributed development. For this project I mined data from large software repositories using C# (on ASP.NET platform) and SQL, provided descriptive statistical analysis of the data and presented my findings in the form of an academic paper.