At Amdox Technologies, I developed financial fraud detection and customer churn prediction solutions using Python and machine learning. I evaluated churn models with accuracy, precision, recall, F1-score, and ROC analysis.
I worked with transaction and login-risk datasets, performing data cleaning, exploratory analysis, and feature engineering. I stored processed records in SQLite and developed a Streamlit prediction app.
I created Power BI dashboards for executive KPIs, fraud analysis, customer analytics, segmentation, and financial analysis. My projects also include a retail analytics platform and dashboards for e-commerce and marketing analysis.
I built a Flask chatbot that combines rule-based intent recognition with responses from a locally hosted Llama 3.2 model via Ollama. It includes a REST backend with conversation memory and a responsive web frontend.
