• Developed a Python-based automation script using powered by Pandas to extract and cleanse 1M+ rows of data across multiple tables, improving data quality and reducing manual ETL effort
• Optimized internal service scripts by refactoring logic and reducing redundancy, cutting processing time by ~60%
• Developed modular pipelines to extract structured data from Databricks, SQL databases, and Excel files on SharePoint, maintaining compatibility across 20+ code files.

• Developed and refactored core backend modules in a Django-based CRM used by 10+ enterprise clients (Levi’s, Burger King, Popeyes), improving code quality and reducing technical debt
• Developed and optimized RESTful API endpoints using Django REST Framework, handling 50K+ weekly requests with 99.9% uptime
• Engineered efficient ORM logic and optimized query performance, reducing DB query count per request by ~40% and improving response times by up to 20% on key endpoints

• Engineered a microservice-based ML framework using Python, Django, and Docker, enabling prediction, clustering, and forecasting features on a web platform — resulted in a 50% reduction in manual analysis time for internal teams
• Analyzed and processed over 200,000 rows of transactional and behavioral data using Pandas and NumPy, generating high-quality datasets that led to a 10% improvement in model accuracy
• Collaborated cross-functionally with backend developers and data scientists to ensure seamless integration of ML services into production, contributing to a faster go-to-market for the platform

• Selected as one of 80 participants in the highly competitive Tokopedia Devcamp Backend Curriculum, chosen from over 5,000 applicants for a 7-day intensive program.
• Proposed and developed the “Tokopedia Recycler System” during the hackathon, implementing the idea into an iOS application using Django REST API and Swift.

Static web app commissioned by a medical intern to provide tuberculosis education and dosage calculators for public health use.

• Built with Nuxt.js and Tailwind CSS, featuring responsive design and interactive drug dose calculators for children and adults
• Developed in collaboration with healthcare professionals to ensure accurate, user-friendly educational content
• Deployed on Netlify, delivering a fast and accessible experience for both mobile and desktop users

Discord chatbot built with discord.py, designed to handle asynchronous user interactions and automate server tasks.

• Handled 1,000+ monthly requests in a community server of 17,000+ members, supporting commands, event triggers, and user data tracking
• Integrated with Supabase (PostgreSQL) to manage persistent user data with scalable, reliable storage
• Built with asynchronous Python to ensure responsiveness under concurrent user activity
• Deployed and maintained the bot to ensure 99.9% uptime and smooth operation across a large user base

Web platform that empowers non-technical users to build predictive and forecasting models for product sales.

• Built a microservice-based ML framework using scikit-learn for model training and forecasting, with a React-based dashboard for user interaction
• Designed and implemented backend logic to support multiple model types (e.g., linear regression, decision trees) for custom training workflows
• Enabled users to upload and test datasets through an intuitive interface and manage their models via a centralized dashboard
• Integrated seamless communication between frontend, backend, and ML services for real-time model management

Interactive web app simulating a real-life Constraint Satisfaction Problem (CSP) scenario for physical distancing.

• Built with Django and React, enabling dynamic simulations of how many people can safely fit in a room under distancing constraints
• Implemented core CSP logic to handle spatial rules, room capacity, and placement validation in real time
• Designed intuitive UI to visualize people placement and constraint violations, enhancing AI learning for students