番茄模型运维（模型开发方向）运营-CQC
Core
Develop and optimize large language models (LLMs) including pre-training, SFT, and RLHF/DPO; build RAG systems and LLM toolchains for business applications.
Role type
Senior IC machine-learning engineer (LLM & RAG)
Builds
Production LLMs, RAG systems, and AI toolchains for business scenarios
Domain
Artificial Intelligence / Large Language Models
Deliverable
production ML models
Required skills
Python, PyTorch, LLM architecture knowledge (Llama, Qwen, ChatGLM, Mistral), SFT/RLHF training, distributed training frameworks (DeepSpeed, Megatron-LM, FSDP), LangChain, LlamaIndex, vector databases (Milvus, FAISS, Chroma), Elasticsearch
Preferred skills
C++/Go, Linux, English technical reading
Responsibilities
Lead LLM pre-training, SFT, and RLHF/DPO; build and optimize RAG systems including document parsing and reranking; develop LLM toolchains and function calling capabilities; construct evaluation and data pipelines for model performance.
