한국에서 꿈을 이루세요.

지금 바로 비잡과 함께 시작하세요.

[NAVER Co., Ltd.] Hiring Research Intern for Voice Translation Quality Evaluation / Multimodal Translation

네이버(주) · 경기 성남시 분당구 정자일로 95

✨Job Description - Voice translation quality evaluation and multimodal translation research - Project 1: Meta-evaluation of a voice translation LLM judge, reliability verification by language pair, and bias analysis - Build a meta-evaluation set based on human evaluation results - Measure the agreement between judge scores and human evaluation scores by language, noise, and genre - Analyze over- and under-evaluation biases by specific types and experiment with mitigation methods - Project 2: Build a screen-dependent mistranslation case set and measure the effect of using screen information - Collect mistranslation cases such as player names, team names, and skill names in eSports live broadcasts - Select cases where accurate translation is possible by using screen information and build a case set by type - Measure the reduction in mistranslations when text on the screen is read with OCR and reflected in the translation prompt - Analyze the effect by sport and screen layout, compared with direct video input methods ✅Qualifications - Experience in simple data processing or writing experimental code using Python - Interest in at least one of the following: machine translation, LLM evaluation, multimodal AI - Ability to read English papers or technical documents and summarize the main points - Must be able to conduct hands-on practice during the internship period - Ability to participate in a full-time experiential internship, 5 days a week, 8 hours a day 💖Preferred - Someone comfortable with systematically organizing experimental results and actively sharing them with colleagues - Experience designing LLM prompts or evaluating outputs using LLMs - Experience participating in projects or research related to speech recognition or machine translation - Experience extracting text from video/images (OCR) or working with vision-language models (VLMs) - Experience applying basic statistical techniques such as correlation to experimental analysis - Familiarity with eSports titles and broadcast screen layouts - Experience publishing papers in related fields or participating in academic workshops/competitions - Persons eligible for employment protection, such as national merit recipients and people with disabilities 🏠Workplace - 95 Jeongjail-ro, Bundang-gu, Seongnam-si, Gyeonggi ⚡️Selection Process - Document screening - Coding test - Job interview - Internship acceptance - Internship performance 👉Recruitment Period/Method - Application period: until 2026.11.13 (Fri) 18:00 - Application method: Apply via the website 📋Other - Internship period: 2026.12.14 (Mon) ~ 2027.03.05 (Fri) - Results for each stage will be notified by email, SMS, and the NAVER Careers (My Page) message inbox - This internship is conducted as an experiential internship - The selection process may change depending on the schedule and situation, and additional steps may be added based on the results of each stage - It may close early once hiring is complete, and the recruitment period may be extended if necessary - If there is any false information in the application, the acceptance may be canceled - If the applicant is confirmed to be subject to international sanctions or related parties, concluding an employment contract may be restricted

Working Conditions

Work Period
4 months
Work Time
10:00 ~ 19:00

Recruitment Conditions

Application Deadline
Closed
Employment type
Intern
Salary
A decision after an interview

Qualifications

Experienced
Entry-Level
Education
No Educational Requirement
Preferred Conditions
Someone who is accustomed to systematically organizing experimental results and actively sharing them with colleagues Experience designing LLM prompts or evaluating outputs using LLMs Experience participating in projects or research related to speech recognition or machine translation Experience extracting text from video/images (OCR) or working with vision-language models (VLMs) Experience applying basic statistical techniques such as correlation to experimental analysis Someone familiar with eSports titles and broadcast screen composition Experience publishing papers in related fields or participating in academic workshops/competitions Individuals eligible for employment protection, such as national merit recipients and persons with disabilities