I'm a Computer Science undergraduate at the University of California, Irvine, focused on AI agent systems and developer tools. Previously a front-end software engineer intern at Taboola.
Right now I'm building Lantern, a local heartbeat daemon that watches a code repository, files issues as tickets, and hands fixes to AI agents. On a schedule, it scans logs, bug reports, and test results, files what it finds as tickets, and fixes what’s safe, leaving anything risky for a person. Every fix arrives as a draft pull request, built in its own sandbox, where agents are never allowed to push to main. Unlike most AI coding agents, it keeps persistent memory in a SQLite database, so it can learn from past fixes and skip issues that aren’t worth a developer’s time.
I wrote about how it works in Lantern: an agent that keeps watch over your repository.