I work at the intersection of two crafts: engineering software and judging AI. At Outlier AI I audit quality across coding, reasoning, instruction-following and conversational AI projects as a Queue Manager Auditor, tracing recurring failure patterns to root causes and turning them into feedback that measurably improves what frontier models learn.

Before that (and in parallel, I'm still a B.Tech student at Bennett University), I built real systems for real users: a placement-data platform that cut manual reporting work by over 70%, the public website and internal CMS/CRM for Shiksha Vijay (a Times Group initiative), e-commerce workflows for BennettSwag, and WordPress learning platforms early on at Cybrot Academy. At 17 I founded HacWar and took an MVP from design to deployment.

My home turf is the terminal, Linux, Arch-flavored, heavily riced. I care about how systems work under the hood: I've built network-security tooling (DNS-tunneling detection on Spark, a wifi deauther, an anti-keylogger), trained and fine-tuned CNNs for medical imaging, and poked at LLVM and computer architecture for fun. I've also done 4+ paid bug-bounty and security-testing engagements.

What ties it together is product judgment: I've delivered 5+ paid freelance projects from requirements to production, led tech for Google Developer Groups on campus, and I treat quality as a discipline, in code review, in AI evaluation, and in the pixels of interfaces like this one.