Thirty years ago, Myerson provided a closed-form solution to the following auction design problem. How to design a single-item auction such that if all the bidders adopted the bidding strategies in their best interest, the auctioneer who has some stochastic knowledge about the highest potential value offered by the bidders on a per item-basis, could maximize revenue. A natural extension to this problem is devising the best strategy to organize an auction in which bidders are competing for multiple items. Daskalakis et al. recently presented an algorithm in order to find the optimal multi-item auction or namely multi-dimensional mechanisms. The goal of this project is to implement this algorithm and analyze and study the optimal achievable strategy through case studies that involve several small well-understood structured auctions. This will further allow the extraction of potential economics-driven interpretations of the algorithm.

I worked at Facebook on Graph Search Infrastructure to improve index servers reliability. I worked with Professor Chris Schmandt at the MIT Media Lab on indoor location sensing using geomagnetic field.