- Built interactive web-based application (JavaScript, NodeJS) with an pattern recognition algorithm to support troubleshooting and generate reports for engineering teams- Deployed with Google App Engine hosting microservices to enhance the reliability of the service sites- Performed feature extraction and optimized regression/classification models (Python, MatLab) to predict uplink rate based on geographic location and achieved 91% accuracy - Used ETL process (Python, Pandas) for AERO connectivity data to perform optimization for network performance

- Applied LSTM(RNN), LeNet-5, CNN in TensforFlow to predict bearings’ temperature for the Electric Power Plant company for maintenance- Detected abnormal data using Decision Tree, Isolation Forest (Pandas, Sklearn)- Validated and etracted grammar and syntax errors (Pandas, Numpy) for data integrity and added log tracing to locate the problematic data - Compiled a customized-error dictionary containing errors and suggestions to inform users failure details - Unit tested (Pytest) project's methods for site's robustness and reliability - Participated in database connection and maintenance and the website set-up