I find what is slow or broken in large production systems, fix it, and build the processes that keep it fixed.
For seven of my nearly nine years at Meta, I owned Instagram Android performance and reliability: incident response, crash and memory work, and agents that debug regressions.
Proof
- StableLed the Instagram-wide crash workstream that halved the crash rate; null-pointer exceptions fell about 3×.
- OperableFounded the Reliability Captain programme and trained 12 engineers to run incidents. It still runs.
- AutomatedBuilt AI agents that investigate and resolve regressions at scale, cutting the supervision load on my team.