Expert Answer
Anonymous
As we approached RTL freeze, there were many pending deliverables along with ongoing regression debugging. I proposed a prioritization strategy to focus on new feature failures in regressions over legacy issues, as the legacy features had already been silicon-proven in previous tapeouts without reported bugs. My goal was to optimize debugging efforts and ensure that critical new functionality was verified before the RTL freeze deadline. I proposed that we focus first on new feature regressions, since legacy features had already been running in production.
- However, my proposal wasn’t well received because stakeholders felt that untested scenarios could still expose latent bugs in legacy features, making it risky to de-prioritize them.
- Instead of pushing my approach, I collaborated with cross-functional teams to gather data on which legacy features had already been verified by other teams (such as validation or software teams).
- Using this data, I presented a compromise strategy:
- Features already tested by other teams were lowered in priority.
- Unverified legacy features remained in scope alongside new feature debugging.
- This data-backed compromise was accepted by management, ensuring new feature verification remained a priority while addressing concerns about legacy coverage.
- We met the RTL freeze deadline without compromising verification quality.
- This experience strengthened my ability to handle pushback constructively and use data-driven reasoning to drive alignment across teams.
