Situation
The tech lead of a system generating fraud signals that powered defenses to prevent 500m$ fraud in Capital One left for new opportunities at the same time re-organizations moved ownership of the system to a different management hierarchy. The system was creating a lot of RTE workload, and its health was preventing development of new fraud defenses
Task
I took the role of principal engineer for Fraud detection and resolution in September (100eng) and one of the first things that I was asked by my manager was to to look into this problem. A significant challenge was that the new tech lead didn’t have some of the necessary skills to execute and also the manager owning that component didn’t have a lot of experience in “data engineering intensive applications”
Action
I worked on two aspects. I identified the knowledge gap, crafted a learning path, and tracked the progress for the tech lead, while working in pair to address some of the most critical issues. Effectively we paired multiple times per week, going through the code, writing code together, while he was catching up in 2 months on some fundamental knowledge. Since we also had headcount open in the group, I worked with managers to understand if hiring someone else with a strong data engineering experience and transferring the tech lead to a more traditional backend engineering job was feasible. I did this because although I think upskilling is the favorite option, I didn’t have a long running relationship with this person and so I hadn’t got any sense of whether he could ramp up so quickly
Result
The tech lead was able to ramp-up quickly, and the system was back on track, reducing RTE of 30%. We were also able to develop some new defenses and get additional 25million$ in saving. Upskilling on data engineering an engineer that had significant fraud expertise turned out to be an effective approach, rather than hiring a data engineer that had no expertise in fraud