2027AIDU-大模型算法工程师(J99938)
Core
Develop and optimize large language models (LLMs) for diverse business scenarios including search, recommendation, dialogue, AIGC, and autonomous driving.
Role type
Senior IC large language model algorithm engineer
Builds
Production LLMs and inference frameworks for search, recommendation, and autonomous driving applications
Domain
Artificial Intelligence / Large Language Models
Deliverable
production ML models
Required skills
Transformer architecture, SFT, RLHF, distributed training, Python, PyTorch, model quantization, speculative decoding
Preferred skills
MoE sparse structures, Latent Attention, Megatron, DeepSpeed, NeurIPS/ICML publications
Technologies
PyTorch, TensorFlow, PaddlePaddle, Megatron, DeepSpeed, Hadoop, Spark, Eagle, DFlash, DSpark
Responsibilities
Design and implement distributed training and inference frameworks; Optimize model structures like MoE and Latent Attention; Support LLM application in search, recommendation, and autonomous driving; Develop quantization and speculative decoding strategies.
Seniority
Senior, hands-on IC