Here’s what I need to start a machine translation post-editing (MTPE) project:
  • Source files in the required delivery format
  • Job instructions, including the target language and schedule
  • End-client requirements, including a style guide and other instructions (if any)
  • Optional linguistic support assets, such as a relevant TM and/or bilingual glossary or termbase, to train the AI for better output. If these assets are unavailable, we can create them from a glossary or sample translation finalized by a human translator. We can also skip the up-front training and refine the AI’s output as we go.
Step 1 (standard turnaround: overnight) – I prepare the text, train the AI, and produce an AI translation. I then provide you with a bilingual Word or .xliff file containing both the source text and AI translation.
Step 2 – The translator post-edits and returns the bilingual file.
Step 3 (standard turnaround: overnight) – Once I receive the post-edited file, I update it in the CAT tool, finalize the text, and send you the final translation in your required delivery format.

 

* For larger volumes, we work on a rolling basis. This allows us to start immediately and continually refine the AI translation for better quality as the project progresses.
* Because I don’t perform manual formatting, you may need to add a post-layout proofreading step before delivering the job to your client.