As a journalist, I had the opportunity to interview Duncan, a former member of Uber’s Marketplace team. During our conversation, he shared with me the challenges and crises that his team often faced while managing core machine learning products that directly impacted the company’s revenue.
Duncan’s team was responsible for products such as targeted promotions, surge pricing, driver incentives, ETAs, pool matching, upfront rider fares, and subscription upsells. The pressure was always on because they knew that if any of these products were malfunctioning, it could lead to the downfall of the entire business.
Every week brought a new potential Data Science Disaster for Duncan’s team to navigate. It typically started with someone noticing something unusual in the data, like a spike in internal metrics signaling a decrease in reliability or an excess of promotional offers being deployed. Each time this happened, they would go into panic mode, fearing the worst and scrambling to find a solution before it escalated.
Sometimes, the source of the crisis would be a tweet from a celebrity who noticed a steep fare on the app, only to see it drop drastically after taking a few steps. These instances would quickly go viral, stirring up negative publicity and putting even more pressure on Duncan’s team to address the issue promptly.
Despite these constant challenges and crises that Duncan and his team faced while working at Uber’s Marketplace team
