Bravo Lucy is an analytics start-up by Nils Rudi, a reputed faculty in Operations Research at INSEAD Business school. We work on a number of interesting problems ranging from Supply chain management to football analytic.
As a core member of the supply chain team, I have been working on developing software that allows retailers to optimally order new products that are to be launched. As there is no historical data available for the demand of such items, the software takes inputs from domain experts(a variant of the Delphi method). As we focus on products with a relatively small life time (cosmetics or fashion), the base model used to decide the order quantity is an extension of the News Vendor model. We also allow the retailer to compare various strategies like ordering components that can be assembled on demand and ordering a group of items with common components together as it allows them to hedge their risk. I have had the opportunity and pleasure to work on these models which has helped me realize some of the mathematical and programming challenges these problems pose.
I have also been involved in developing a model that tries to forecast the demand for a new product from the historical data that is available from certain "like items" (items that are similar to the new product in a certain sense and for which we have historical data available) .The algorithm embeds the items in the space of their attributes (say color, variant,texture) and uses inputs from each like item, weighted by its distance from the new product in this space, to forecast the demand.
Cloudinfra: Work on recommendation systems and search. (To be updated)
BicycleAI: Work on task oriented conversational agents (To be updated)
ML Consultant: Multiple projects (To be updated)
Apollo.io (To be updated)