I've built an end-to-end customer churn prediction platform using Python, XGBoost, and automated data preprocessing pipelines.
I integrated the trained model into Streamlit for risk prediction, visualization, and automated reporting, while analyzing features associated with customer churn.
In my research on misinformation detection, I developed a CLIP-based image-text embedding system and designed workflows for preprocessing, fine-tuning, evaluation, and performance analysis.
Through an industrial attachment at Brain Station 23, I've worked with Python-based ML workflows and gained exposure to ML pipelines, AI deployment practices, and software development processes.
