Piotr Keller’s experience includes roles as program manager and strategist for language industry heavyweights Lionbridge and Lilt. Here, he describes what attracted him to the localization field and how he sees it evolving in the future.

Why do you enjoy reading MultiLingual magazine?

There’s always something interesting about the technology in the language industry — lots of perspectives.

How did you get involved in the translation business?

After nearly 10 years of teaching English as a foreign language, I decided I needed a change. I joined a language team as a quality checker, and it was the best start one could hope for. I jumped directly into software localization and got to work hand in hand with some of the most knowledgeable and skilled language teams in the company. Deciphering placeholders and variables, running linguistic and functional testings, and working with developers on the apps got me hooked. After a few years, my focus shifted to team leading, managing production and quality for client programs.

Since you entered the translation industry, how has the business landscape changed?

The last 10 years brought in quite a bit of change. I remember the move from SharePoint and File Transfer Protocol (FTP) to the very first translation management systems (TMSs), which gradually evolved to genuinely support cross-team collaboration. From there, we transitioned to online computer-assisted translation (CAT) tools, transforming the way linguists, project managers, and quality assurance (QA) professionals worked together. Then came the machine translation (MT) revolution, where we witnessed the evolution from statistical to neural MT, bringing with it significant improvements in fluency and contextual accuracy. Finally, the advent of large language models (LLMs) took things further, enabling even more nuanced and sophisticated language processing.

Through all of this, one thing remained constant: the industry’s goals of speeding up translation work, optimizing costs, and maintaining quality to meet the ever-rising bar for “good.” Each new tool and system wasn’t just a technical upgrade; it reshaped the way we look at metrics and redefined how we measure and define language quality.

Could you share your experience working with your first client or on your first project?

My first client was a major software development company. I was responsible for the linguistic quality of a set of their desktop applications and (later on) their mobile apps. The first project I really remember was a linguistic testing, bug regression, and corrections project across almost 30 languages. This was my first encounter with a bug-tracking software, hundreds of screenshots coordinated across three continents. This is what I learned:

• Good documentation is key.
• Shorter translation is not the answer for layout issues.
• Time spent on planning the work is never time wasted.
• The working day really ends around 5 p.m. Pacific Standard Time on a Friday.

Do you believe it’s a good time to enter the translation business?

The answer to this one is tricky. From the point of view of the linguist, you need to come to grips with the fact that your standard translation tasks will most likely be post editing. As LLMs develop, this will likely move over to marketing and the trick will no longer be ensuring fluency as much as accuracy — or simply making the text sound more human. I firmly believe that a good linguist is worth their weight in gold. And while the industry moves forward at an astonishing rate, the need for capable managers will not be replaced anytime soon either. If you have the knack, absolutely!

Where do you see yourself professionally in the next 10 years?

Ten years in the future is definitely too far out to see myself anywhere or even have an outline of where I’ll end up. What do I see myself doing? Working with amazing people on something worthwhile.

What predictions do you have for the future of the translation industry?

That one’s easy. LLMs will evolve to handle more complicated content types. With luck, LSPs will use it to enhance the work; with less luck, they will provide more usable texts for dry post editing. Self QA will be next in line, and once the basics are nailed down, less false-positive glossary checks and style guides will follow. Next, probably review and arbitration.