One of the research topics that fascinates me is how we make decisions. I myself felt very confused by how supposedly easy it was to find activities near me, and yet how paralyzingly hard it was to choose. The paradox of choice, decision fatigue, and numerous decision biases lead to increasing stress and a burgeoning dissatisfaction with monotonous lives. Instead of addressing these complications, common recommendation platforms return overwhelmingly crowded and misleading results, only complicating the decision-making process.

This was the motivation to found my startup, Shuflix, a completely new type of search engine, that finds fun things to do right here and now, fundamentally challenging not only the assumptions outlined above, but also the very nature of what online search means.

Unlike other search engines, Shuflix lacks a text box where you type in keywords, its results are in no discernible order, and you can only see a handful of them at a time. The project was shaped by research into the fields of psychology and behavioral economics and verified with surveys in major East Coast cities and user behavior collected worldwide.

When a user searches for something to do, Shuflix’s algorithm draws from millions of events and places around the world. It then filters, personalizes and shuffles the results using a distributed No-SQL database, trainable weights, and biases for personalization, diverse sampling techniques and K-modes clustering for balancing the shuffled results. It then returns a small number of diverse results that comply with the science on effective decision-making. If no result in the short list catches the users’ imagination, a new list can be generated instantly.