Sinan Mouline

I worked on a dashboard to allow customers to launch advertising campaigns targeting several social media platforms.

The first section collected general information on the campaign such as the name, the budget, the bidding type, the duration, and so on.

The second section allowed customers to specify target audiences and creatives. The customer could create several combinations of an audience with creatives, and we allowed for saving target audience groups, creatives, and combinations as templates to be reused.

The third section then verifies that the obtained information is correct. It also provides feedback related to the budget and campaign duration to let the customer know if the campaign will be effective.

I also worked on building the event log, which described all the steps taken by a user on the platform.

Say Say K has a mobile application to help people learn languages, through watching their favorite movies and TV shows.

I worked on the content management system to enable the content team to manage the movies and TV shows displayed in the application.

I started the project by writing command-line tools in Python interfacing with AWS. The content was versioned to allow for rollbacks if necessary and distinguish changes over time.

One feature was to convert the SRT subtitles into a proprietary format that contained the subtitles in both the learning language and the user's native language, by calling the Google Translate API recursively. The content team would then improve the translation further in another iteration.

Another feature was the ability to copy restore the state of the content from a previous version.

Later on, I built a UI with JavaScript and Vue.js to make the project more user-friendly. The UI interfaced with a back end written in Python and Flask.

Once Say Say K started to have media partners, we needed to connect to their MRSS feed to be notified of new available content and process it on our end.

I built a script to listen to the MRSS feeds for all our partners and start the processing when new content is detected. I used Python and Redis, which allowed us to spin up as many instances as desired to perform a distributed processing in parallel, which saves time.