Uber’s Marketplace team, led by Duncan Gilchrist, was under constant pressure to ensure that their core machine learning products were running smoothly. With billions of company dollars at stake, even the slightest issue could potentially sink the entire business. The team was always on high alert, ready to tackle any crisis that came their way.
Every week brought a new potential Data Science Disaster for the team. It would typically start with someone noticing something suspicious in the data, such as a spike in internal metrics like “zeros” indicating that riders were opening the app but not finding any available cars. This raised questions about the reliability of the service and whether they were deploying too many promotions or the wrong incentives for drivers.
Sometimes, a crisis would be triggered by a tweet, such as a celebrity complaining about a high fare that drastically changed after walking a short distance. These incidents would quickly go viral, sparking public outrage and causing even more chaos for the team. Despite these challenges and pressures, Duncan and his team remained determined to address each crisis head-on and find solutions to prevent them from happening again. Their dedication to analyzing data, identifying issues, and implementing improvements helped them navigate the turbulent waters of the marketplace and keep Uber running smoothly.
