Swaapnika Motapothula

Swaapnika Motapothula helped build an AI teammate for on-call engineers at Meta with Chinmay Gandhi, Vlad Tsvang and their colleagues. Their team gave each on-call rotation a shared agent that starts investigating when an alert fires. It can connect an incident ticket to a recent deployment, check how engineers resolved similar incidents and have a proposed fix ready when the engineer on duty joins the investigation.

Swaapnika studied computer science and engineering at the National Institute of Technology Warangal, where she earned a campus placement with JPMorgan Chase. She worked as a teaching assistant at Arizona State University and earned a master’s in computer science there in 2022 before joining Meta as a software engineer. In 2025 she returned to the building where she had taught for an ACM student workshop with Himanshu Shekhar. Swaapnika discussed how engineers judge an incident’s severity and how a team can review what happened afterward without assigning blame.

Swaapnika explains her team’s autonomy ladder by comparing it to training a new engineer on call. The agent first investigates, then drafts fixes for review. The team grants permission to act independently for a particular task only when the agent has a track record doing it, a mistake would cost little and engineers can catch it easily. That lets the agent close a duplicate alert on its own. If checkout starts failing after a change to the payment client’s timeout, the agent can show the engineer when the errors began and draft a rollback. The engineer decides whether to run it.