It was a hugely ambiguous space and our team consolidated a number of complex, pre-existing solutions, and collaborated with other teams to determine how engineers could reason about platform-specific resources in a more approachable & scalable way. One key aspect was the interviewing of engineers from a wide selection of the different parts of Meta's infrastructure, to determine what abstractions could be used to make the management of resources more understandable. This was particularly important in avoiding potential errors, which risked costing the business huge sums of money.