LG AI Research

Builds AI models for research and industrial use

Large Language Model Research and Development Intern

Natural Language Processing & LLMsConfirmed live in the last 24 hours

InternshipSeoul, South KoreaNo salary listedIn-person

About the Job

Requirements

Applicants should be expected undergraduate graduates, already graduated, or currently enrolled in a master's or doctoral program.
Experience developing deep learning and machine learning algorithms is required.
Proficiency with Linux, Python, PyTorch, and Hugging Face language models such as BERT, T5, and GPT is required.
Applicants should have a major in Computer Engineering, Computer Science, Artificial Intelligence, or Machine Learning.

Responsibilities

Conduct research, development, and experiments related to large language models.
Build pipelines for collecting and preprocessing training data.
Measure benchmark performance for in-context learning and fine-tuning.
Develop prompt engineering for in-context learning.
Optimize inference engines.
Research and develop multi-task instruction tuning and methods to improve performance.
Research and develop toxic classifiers and safety methods for large language models.
Research methods to improve language-model reinforcement-learning performance and develop distributed training.
Research and develop model-compression methods for large language models.

Desired Qualifications

Experience developing language-model pretraining is preferred.
Experience developing with DeepSpeed and FastTransformer is preferred.
High leaderboard rankings, competition awards, or open-source contributions are preferred.
Experience with the machine-learning production cycle, including data collection, model development, model testing, and serving, is preferred.
Experience developing with large-scale data-processing frameworks such as Apache Beam, Amazon EMR, or GCP Dataproc is preferred.
Experience publishing papers related to language models is preferred.

<팀 소개>

EXAONE Lab은 초거대 언어 모델과 관련된 기술을 선행 연구/개발하는 조직입니다.

초거대 AI 최첨단 기술을 고도화하고 이를 실제 서비스에 적용시킴으로써 세상을 더욱 이롭게 바꿔 나갈 열정적인 구성원을 찾고 있습니다.

<수행 업무>

초거대 언어 모델 관련 연구/개발 및 실험

- 학습 데이터 수집 및 전처리 파이프라인 구축

- In-context learning 및 fine-tuning 벤치마크 성능 측정

- In-context learning을 위한 prompt engineering

- Inference 엔진 최적화

- Multi-task instruction tuning 및 성능 향상 기법 연구/개발

- Toxic classifier 연구/개발 및 초거대 모델 safety 방법론 연구/개발

- 언어 모델 강화학습 성능 향상 방법론 연구 및 분산 학습 개발

- 초거대 모델 경량화 (model compression) 방법론 연구/개발

<지원 자격>

- 학부 졸업 예정자 및 기 졸업하신 분, 석사/박사 재학중이신 분

- Deep learning, machine learning 알고리즘 개발 경험 필수

- Linux, Python, PyTorch, HuggingFace 언어 모델(BERT, T5, GPT 등) 활용 능력 필수

<우대 사항>

언어 모델 사전 학습 개발 경험 우대

DeepSpeed, FastTransformer 개발 경험 우대

Leaderboard 상위권 달성, competition 입상, 오픈소스 활동 우대

ML production cycle (데이터 수집, 모델 개발, 모델 테스트, 서빙) 경험 우대

대용량 데이터 처리 framework (Apache Beam, Amazon EMR, GCP Dataproc 등) 개발 경험 우대

언어 모델 관련 논문 게재 경험 우대

<전공>

컴퓨터공학, 전산학, 인공지능 등 머신러닝 전공자

About the Company

LG AI Research

Builds AI models for research and industrial use

LG AI Research develops AI models and industry-focused systems for businesses and researchers as LG Group's central AI research organization. Its EXAONE family handles language and images, while specialized platforms address data analysis, materials discovery and drug research. The institute turns research into tools for company documents and operational decisions, working with LG affiliates and outside partners to apply models to specific business problems. It also distributes model weights to developers and supports commercial applications through industry collaborations and licensing. Established in 2020, LG AI Research is headquartered in Seoul, South Korea.

Company size: 201-500
Company stage: Not listed
Total funding: Not listed
Headquarters: Seoul, South Korea
Founded: 2020

Recently Posted Jobs

Explore more opportunities at LG AI Research

See more jobs