算法工程师-数据智能（杭州）
Core
Optimize data synthesis algorithms for pre-training/SFT/RLHF stages and lead R&D of large language models (Code-LLM, logic reasoning) to solve technical challenges in NL2Code and complex reasoning chain generation.
Role type
Senior IC large language model researcher and engineer
Builds
Production large language models (Code-LLM, reasoning models), RAG-QA robots, and data insight robots
Domain
Artificial Intelligence / Large Language Models / Data Intelligence
Deliverable
production ML models
Required skills
Deep learning training frameworks (PyTorch, Huggingface), Large language model architectures, Embedding architectures, Post-training techniques, NL2Code generation, Complex reasoning chain generation, Test-Time Compute, Parameter-Efficient Finetuning, Scaling law analysis, Data synthesis algorithms
Preferred skills
Top-tier conference publications (ACL, NeurIPS, ICML, EMNLP, ICLR), Competitive programming awards
Technologies
PyTorch, Huggingface
Responsibilities
Optimize data synthesis algorithms for pre-training, SFT, and RLHF; Lead training innovation for Code-LLM and logic reasoning models; Deploy LLM technologies in real-world scenarios (RAG-QA, data insight robots); Explore open-source SOTA models and implement post-training for data intelligence.
Seniority
Senior, hands-on IC
