Experience

Norna

• Finland
• Personal Care Product Manufacturing

• Data Engineer

  • Aug 2023 - Present

  Stockholm, Stockholm County, Sweden Managing over 10,000 products to deliver optimum data analytics to customers. Building more than 20 Python-based scripts for MongoDB data management automation. Tackled more than 10 data management issues each week to deliver clean and accurate data analysis. Technologies include Python, MongoDB, Selenium, and Linux



Telkom University

• Indonesia
• Education Administration Programs
• 700 & Above Employee

• Machine Learning Research Assistant

  • Aug 2022 - Present

  Bandung, West Java, Indonesia Successfully implemented NearCount and PSNB, prototype selection algorithms, on over 20 kinds of datasets to later measure the confidence score using LMPNN, LMKNN, KNN and PNNR algorithm Implemented NearCount based on the corresponding paper. Create and modify PNNR, LMKNN and LMPNN as callable Python function and fit it with the SHAP algorithm. Currently working on visualizing data to explain its feature contribution towards prediction results using SHAP and LIME.



PT Infrastruktur Telekomunikasi Indonesia (Telkominfra)

• Telecommunications
• 200 - 300 Employee

• Full Stack Software Developer

  • Jun 2022 - Aug 2023

  Jakarta, Indonesia Developing user interface, front-end, back-end and documentation of project, and network monitoring dashboard using Django, JavaScript and SQL Server to provide real-time analytics and forecasting to increase network KPI. Built out the data management and automation from the ground up using Python and MongoDB to provide automation to data management. Created data visualization for Area 2 using Python and Tableau. Currently learning and working on web deployment and AWS.



Universitas Telkom

• Indonesia
• Non-profit Organizations
• 1 - 100 Employee

• Machine Learning Research Assistant

  • Aug 2020 - May 2021

  Bandung, Jawa Barat, Indonesia Successfully created the algorithm for a new method to classify data called Multi-Voters Multi-Commission Nearest Neighbor (MVMCNN) that outperformed the previous study, LMPNN, and KNN, with Friedman Mean Rank ratio 1.85:1.92:2.18 in 33 datasets. Successfully implemented Grapheme to Phoneme (G2P) in an Indonesian dataset using an existing transformer with 10% error rates.