Er. Pooja Derashri

I wrote my first program — a clumsy bit of Turbo C that computed prime numbers — as a diploma student, and I still remember the strange thrill of watching a machine do exactly what I told it to. That thrill hasn’t worn off in fifteen years, even as the field has moved from procedural C to distributed systems and neural networks. What has changed is my belief about what a Computer Science classroom is for.

I spent four years as a software engineer building backend systems for a logistics firm in Pune before returning to academia, and that industry stint shapes how I teach today: I care less about students memorizing syntax and more about whether they can reason about trade-offs — why a hash map beats a list here, why a system fails under load there. My courses in Data Structures, Operating Systems, and Machine Learning are built around small, ugly, real problems rather than tidy textbook ones.

My current research sits at the intersection of applied machine learning and network security — specifically, anomaly detection in IoT traffic — and I try to fold every dataset and every messy false-positive I encounter back into my lectures. A professor who has stopped debugging their own code, I think, has little business teaching students how to debug theirs.

— Er. Pooja Derashri