Staff Software Engineer
Core
Design, implement, and optimize MRI image reconstruction algorithms and software, including classical methods and deep learning-based reconstructions, while developing core Agent runtime systems and multi-Agent collaboration mechanisms.
Role type
Staff Software Engineer (LLM & Agent Systems)
Builds
Clinical-grade MRI reconstruction products, LLM evaluation systems, and scalable multi-Agent runtime systems.
Domain
Medical Technology / Artificial Intelligence / Large Language Models
Deliverable
production ML models
Required skills
Large language model fine-tuning (SFT, RLHF), Prompt Engineering, Function Calling, Chain of Thought reasoning, LLM evaluation system design, Agent runtime architecture, Python programming, Deep learning frameworks (PyTorch, TensorFlow), Parameter-efficient fine-tuning (PEFT, LoRA)
Preferred skills
CUDA/CuFFT/CuBLAS, OpenMP, SIMD vectorization, Git/Jenkins/DevOps, Open-source contributions
Technologies
LLaMA, Qwen, ChatGLM, LangChain, AutoGPT, OpenCWA, PyTorch, TensorFlow, Python
Responsibilities
Fine-tune large language models for vertical scenarios; Design high-efficiency prompt templates; Optimize Function Calling and tool invocation; Implement advanced reasoning technologies (CoT, reflection, long context); Build comprehensive LLM evaluation systems; Develop core Agent runtime systems and multi-Agent collaboration mechanisms; Connect and extend open-source Agent frameworks; Collaborate with product and engineering teams to translate research into solutions.
Seniority
Staff, hands-on IC with strategic influence
