At SCIOBYTE (OPC) Pvt. Ltd., I architect end-to-end ML pipelines in Python and create reproducible EDA dashboards that reduced manual analysis effort by 40%. I also trained and evaluated classifiers that achieved 85% accuracy on a binary prediction task.
I implemented a generative AI pipeline using Azure OpenCV transformations to synthesize labeled image samples, expanding the training set by 20% and improving downstream model recall by 5%.
At Dhee Coding Lab, I engineered feature pipelines for three real-world projects and developed classification models that reached up to 82% accuracy. My projects include PhishSafeAI, a phishing URL detection model that achieved 96% accuracy, and a semiconductor fault classification system that achieved 92% accuracy.
