LLM Algorithm Engineer

Responsibilities

• Design distributed deployment solutions based on NVIDIA hardware architecture (NVLink/NVSwitch)
• Lead framework selection and performance tuning for vLLM, TensorRT-LLM, and SGLang
• Build multi-modal GPU cluster management systems to optimize KV Cache storage
• Design Prompt Engineering strategies combined with RAG technology
• Lead model fine-tuning using LoRA/QLoRA and PEFT methods
• Develop enterprise-grade internal toolchains including code assistance and private knowledge systems
• Design external customer systems such as smart customer service platforms and multi-Agent assistants

Requirements

• Master’s degree or above in Computer Science, AI, or related fields
• 3+ years of professional experience in NLP/LLM projects
• Proficiency in PyTorch/TensorFlow and Transformer architecture
• Familiarity with CUDA programming and NVLink topology design
• Mastery of LangChain, vLLM, and SGLang frameworks
• Experience with pre-training or fine-tuning open-source models like LLaMA or DeepSeek

Preferred Qualifications

• Experience in NVIDIA chip operator development and CUDA kernel optimization
• Knowledge of IVR, ACD, and call center technologies
• Experience building enterprise-level code assistance tools
• Experience constructing knowledge graphs in e-commerce or internet industries

Benefits

• 20 days of annual leave (increasing up to 30 days)
• 13 public holidays plus parental, sick, bereavement, and marriage leave
• 22 weeks paid maternity leave and 4 weeks paternity leave with WFH flexibility
• Monthly lunch allowance and onsite gym
• English courses and access to Udemy/LinkedIn Learning