At VIT Chennai, I designed and evaluated machine-learning models for PI-IDS, a protocol-invariant intrusion-detection method that improved AUC by 20% under zero-shot domain transfer. The manuscript is under review.
I co-developed a speech keyword-spotting model for Samsung’s EnnovateX hackathon, which advanced to the second qualification round, and built an MLOps pipeline for network intrusion detection. At Hackfinity, I delivered production React.js components and fixed a critical frontend data-pipeline bug.
