This study investigates how an Indonesian EFL student uses DeepL, a machine translation (MT) tool, as part of her translanguaging practices in academic writing, and how she refines machine-generated texts to meet academic standards. Using a qualitative case study design, this research employed semi-structured interviews, writing assignments, and screen recordings to collect in-depth data. DeepL was specifically chosen among other MT and AI tools due to the participant’s consistent preference, contextual accuracy for academic writing, and a unique alternative-word-suggestion feature that appears to facilitate the participant’s text refinement process directly. The findings suggest that DeepL acts as a learning resource that supports vocabulary development, paraphrasing, and linguistic reflection. The participant critically engaged with DeepL’s translation results by employing several strategies, including back-translation, paraphrasing, and text evaluation, demonstrating an awareness of meaning, tone, and academic style. These practices reflect the translanguaging theory that the use of multilingual repertoires can be supported by digital technology in the construction of meaning. The novelty of this research lies in its rich, contextual insights into collaborative interactions between humans and machines in a single case, thereby providing an exploratory foundation for future, larger-scale comparative studies. The findings of this research also contribute to the field of applied linguistics and EFL pedagogy by proposing the pedagogical integration of MT tools to enhance critical digital literacy and reflective language learning.