7 дней назад
AI Analytical Automation Engineer
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
Текст:
TL;DR
AI Analytical Automation Engineer (LLM/AI): Implementing, deploying, and optimizing enterprise large language model systems for semiconductor operations with an accent on prompt engineering, fine-tuning, RAG, and agent-based applications. Focus on building end-to-end AI solutions, improving inference performance through quantization and distillation, and integrating models with cloud deployment infrastructure.
Location: Chengdu, China
Company
develops integrated power semiconductors and system power delivery architectures.
What you will do
- Implement, deploy, optimize, and maintain enterprise large language model projects.
- Design AI solutions and improve model training and inference efficiency.
- Apply prompt engineering, LoRA and QLoRA fine-tuning, performance evaluation, quantization, and distillation.
- Build AI applications using LangChain, retrieval-augmented generation, and agent frameworks.
- Research new AI technologies and maintain technical expertise.
- Collaborate across departments and complete additional assignments from the supervisor.
Requirements
- Strong foundation in machine learning and deep learning.
- Proficiency in Python, PyTorch, prompt engineering, and large language model fine-tuning.
- Knowledge of Transformer architecture, natural language processing, RAG, and agent-based AI applications.
- Experience with model optimization, Git, Docker, and cloud platform deployment.
- Experience delivering projects end to end, with strong business acumen and cross-department communication skills.
- Good English and Chinese communication skills required.
Culture & Benefits
- Creativity, sustainability, and product innovation are emphasized.
- Equal employment opportunity and a diverse employee population.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →