At Airbus India Private Limited, I build predictive monitoring models from operational time-series data. They reduced unscheduled maintenance by 22% and operational failures by 15%.
I fine-tuned models and tested them before deployment to bring false-positive alerts from 10% to 3%. I also monitored live model performance using AWS SageMaker.
I applied clustering and anomaly detection to historical data to uncover fuel leakage and non-optimal usage patterns, saving 100 tons of fuel. I also built a RAG Agent to help engineers resolve queries 45% faster.
During my Airbus India Private Limited internship, I worked on fault-prediction models and built fault models in Simulink. In a project, I developed a PyTorch attention Encoder-Decoder model to translate regional-language sign boards into English.

