Whenever you are embarking on a new endeavor, it never hurts to test things out first. In a business setting, that holds true as well. Pilot projects are a good way to test the feasibility of a new solution or tool, which allows you to execute it on a smaller scale in controlled settings, but evaluate its scalability as a full-scale initiative.
Here are two pilot projects that I undertook recently to test some translation technology: namely training a custom machine translation (MT) engine as well as evaluating two Translation Management Systems (TMS).
Training a Custom MT Engine for Chinese Drama Subtitles
For this project, we were interested to evaluate the feasibility of training a custom MT engine for Tencent Media, which could be used to machine-translate the subtitles for Chinese dramas that the company produces. We first ran a pilot project to evaluate the initial training results and final output quality of the MT engine, before deciding if we would recommend proceeding with a full-scale project.
Check out our video and final updated proposal below!
Evaluating two TMSs for the Singapore Tourism Board
In this project, I wanted to simulate the process of comparing and evaluating two TMSs for the Singapore Tourism Board (STB), and wanted to find out what kind of organizational needs a government agency like STB would have. It was an insightful project as I was able to go through the analysis and final presentations of findings, just like one would in a real business setting – either as a vendor LSP or in-house client ordering translations.
My two chosen TMSs that I compared were Smartcat and Wordbee – check the video and my final presentation deck to see the results of the evaluation!
