Md Ashfaq Salehin

This personal project is an Android application for recording, editing, and mixing audio files. A limited open-source version of this application is available here: Github.com/ashfaq1701/fast-mixer.

Project Highlights:
• This is an NDK-based project. Most of the operational parts of the project are done as native codes.
• This project is built with Android architecture components, MVVM, LiveData, navigation, data binding, coroutines, and more.
• Hilt is used as a DI framework.
• RxJava and a reactive programming pattern are used to act against various view events.
• Used Google Oboe as the audio streaming framework and FFmpeg for decoding and encoding audio.

A heavy NDK-based Android application for recording, editing, and mixing audio.

This project is done using the latest Android architecture components, MVVM, LiveData, navigation, data binding, coroutines, and more.

Key Deliverables:
• Used Hilt for dependency injection.
• Used RxJava and a reactive programming pattern to act against various view events.
• Designed some complex custom views doing heavy rendering jobs. Used RxJava observers heavily inside those views to react promptly against UI actions.
• Performed all heavy computation and IO tasks in background threads (IO Dispatcher and Default Dispatcher).

Most of the operational parts of the project are done as native codes. The solution is scalable and will allow many layers of further improvements without a significant structural refactor.
The app uses FFmpeg for decoding audio and libsndfile to write into a WAV file. and ASAN and HWASAN to track memory leaks.

I developed this application as part of my job responsibilities while working at OPN. I built many core features of the application with native Kotlin MVVM architecture. and, as part of my job role, I owned some important features of this application.

This project was done using Spark, Hadoop, and Scala. This is a back-end service to calculate a summary of vehicle prices for a particular year, make, model, and trim. This deals with a huge amount of data. Each night it runs and calculates a summary of hundreds of millions of raw sales records and summarizes statistical information grouped by each year, make, model, and trim. As data can be too noisy, the system runs outlier detection on the rows to pre-filter out anomalies. It is a very dynamic back-end service. This is still under development.

This is an ongoing project for my M.Sc. in Artificial Intelligence dissertation. In this project, we process a massive volume of unformatted microservice log data (our data volume is over 15TB). It handles system resources efficiently, processing 15TB of data on a personal computer within just six hours.

After processing the data, we construct windowed graphs (of specific intervals) from this data. Then, we apply dynamic system theories and statistical manipulation of these windowed graphs. We aim to find distinguishable network states from the data by applying mathematical and statistical concepts and techniques. We will later use a graph neural network to streamline the process. The shared code implements this whole pipeline end to end.