Leading the development of an AI-native ERP platform for a fiber-to-the-home (FTTH) client, from requirement discovery and engineering blueprint to release. The system is built as a TypeScript monorepo with NestJS, Prisma ORM, and PostgreSQL on the backend, React Router on the frontend, and an OCR service that extracts data from scanned documents.

Own the technical direction of the project: defining coding and API conventions, maintaining the specification and task backlog, reviewing pull requests, and delegating modules to fellow engineers.

Delivered an ERP implementation for an oil and gas component distributor on ERPNext v16, covering requirement mapping, customization, master data migration, and self-hosted deployment.

Maintained legacy code for previous backend projects, which utilized Express.js for REST APIs and MongoDB with Mongoose ODM. Performed code refactoring to implement DRY (Don't Repeat Yourself) principles for improved maintainability.

Designed and developed the Indotruck Portal, a web-based application for digitalizing documents across various transactions, including asset management for the IT department and clearance sheet management for HR. The system was built using Express.js and MongoDB with Mongoose as well.

Successfully developed a REST API for a face recognition access door lock system using FaceAPI.js and integrated it with third-party hardware devices on the doors.

Successfully migrated an existing full-stack ticketing application to a more modern technology stack for tracking complaints across departments. The updated application was built using the MERN stack (with MySQL as the database) and Sequelize ORM for efficient database management.

Deployed the application to self-hosted server in the local network using Apache HTTP Server.