Code Pair

Pair to pair coding platform for developers to collaborate on coding problems and interview preparation.

GO React Websockets WebRTC

Cryptarithm Solver

a lightweight solver for cryptarithm puzzle that running fully on client.

Svelte Tailwind CSS

Chronodir

Simple cross-compiled CLI tool that creates directories with timestamp prefixes.

C Cmake

Http Server from Scratch

Golang net/http like api, build from scratch, for codecrafters.io challenge.

Go

Puskesmas Satisfaction Analysis

Analyze the citizen satisfaction of each puskesmas in Surabaya, based on Google Maps Review.

Selenium (python) Bi-LSTM + Attention Tableau

Microservices for Bank Loan Use Cases

This project explores the microservices architecture for Account, Card, and Loan services. It is part of a self-paced learning module focused on understanding microservices principles.

Java Springboot Docker Kubernetes Helm

Multilang Paraphrasing SaaS

Simple and easy to use tool for paraphasing text. It uses LLM to generate paraphrases of the input text. The tool is built using NextJs and FastAPI based on T5 model.

Next.Js FastAPI Transformers

BFS Maze Solver

OCR-ing Maze Puzzle, solve it using BFS, and wrapped as GUI Program.

OpenCV PyQT5

Code Plagiarism Detector

Backend service for detecting code plagiarism. The service uses a combination of lexical and semantic analysis to detect similarities between code snippets.

Pygments NLTK FastAPI Lexical Analysis

PinjamLab

A simple web app for borrowing and lending lab equipment. The app is built using Next.Js and TRPC for the backend.

Next.Js Tailwind CSS TRPC Prisma

Digit Recognizer from Scratch

Applying Neural Networks to recognize handwritten digits. The model is trained from scratch using Pytorch and NumPy.

Flask Pytorch Next.js Tailwind CSS

Nesa-Judge

Online Judge system for basic programming assignments, built on top of DOMjudge. The system is containerized using Docker.

Symfony Docker MariaDB

OCR for Plate Numbers

Training a model to recognize plate numbers from images. The model is trained using TROCR and be managed using ML-Ops tools such as Weights & Biases (Wandb).

TROCR OpenCV Wandb