Summary
László Laki is a Senior Natural Language Processing Engineer with a Ph.D. in machine translation and over a decade of experience building production-grade MT and LLM systems across academia and industry. He architected the translator backend at Globalese—covering model training, XLIFF translation, cloud integration and recent ChatGPT augmentations—and introduced innovations like subword-based XML tag replacement, domain adaptation workflows, and GPT-boosted terminology extraction that measurably improved translation quality. In parallel he leads development at the Research Institute for Linguistics, driving Hungarian-centric LLM work including a monolingual 6B GPT3 and a trilingual GPTrio model, plus instruction-tuned variants and LLAMA-index experiments for intelligent search. Now at memoQ, he combines research rigor with product-focused engineering to deploy domain-adapted, scalable NLP features for real-world localization workflows. Notably, his work bridges legacy statistical MT expertise and modern transformer/LLM approaches, making him adept at lifting research models into operational translation pipelines.
10 years of coding experience
13 years of employment as a software developer
MSc, Natural Language Processing, MSc, Natural Language Processing at Katholieke Universiteit Leuven
Doctor of Philosophy (PhD), Information Technology, Doctor of Philosophy (PhD), Information Technology at Pázmány Péter Katolikus Egyetem
English, French, Hungarian