Hi, my name is Safiq Faray
I'm a Developer.

First project with Klugee, a CLIL-based institution. It's a web application built with Laravel. My role is implementing both the front-end (HTML, CSS, JS) and back-end with MySQL as the database. The features include user authentication, user management, and lesson planning. The project was done in 4 months (August 2020 - November 2020) and it is now deprecated.

Second project with Klugee. It's a web application built with Laravel. My role is implementing both the front-end (HTML, CSS, JS) and back-end with MySQL as the database. The features include user authentication, user management, teaching logging, student progress logging, teacher management, and accounting management. The project was done in 6 months (January 2022 - June 2022) and it is still maintained. It is a proprietary software and the source code is not available to the public.

During my time as an intern at PT Javan Cipta Solusi, I was a part of a team that developed an undisclosed government's information system. It's a web application built with Laravel. My role is implementing both the front-end (ReactJS) and back-end with OracleDB as the database. The features include user authentication, user management, and information management at a large scale. I was participating in the project for 3 months (June 2023 - August 2023).

With the same functionality, Klugee Roster is being rebuilt with a new design and new features. The project is built with Laravel 11 and now utilizing it's own SSO system. The main difference is the architecture of the project, which is now using microservices.

I was involved in a vote count detection project for election forms, built in Kotlin as part of an experiment for a college course. Another member and I were responsible for conducting experiments on an Optical Character Recognition model using TensorFlow with a simple CNN architecture. We achieved 98% accuracy on the test set and later 80% accuracy in real use cases.

This is my chosen topic for my final project in the NLP course. The project is about transforming Indonesian informal language into formal language. The project is built with Python and the features include data preprocessing, model training, and model evaluation. The model used is ByT5, a encoder-decoder transformer model. The project was done in 3 months (September 2023 - November 2023).

Multi-label emotion classification on Indonesian text is my chosen topic for my undergrad thesis. Dataset construction is one of the activities that I have done. The annotation of the dataset is done by me and my team (8 people in total). The dataset itself is taken from Lin et al (2022) research of the same topic, but the labelling is done by us. The project was done in 2 months (March 2024 - April 2024) with over 4000 data annotated.

Several projects on image processing desktop apps using matlab/python about image enhancement, restoration, segmentation, and deep learning stuff.