大模型智能标注与智能评测算法工程师（全模态方向）（J105943）
Core
Build and maintain an intelligent annotation and evaluation platform for text, image, audio, and video to support full-modal large model development and deployment.
Role type
Senior IC full-modal algorithm engineer (annotation & evaluation)
Builds
Full-modal data production pipelines, intelligent annotation algorithms, Judge/Reward models, and automated evaluation systems for multimodal content.
Domain
Artificial Intelligence / Large Language Models / Multimodal Systems
Deliverable
production ML models
Required skills
Python, PyTorch, Large Language Models, Vision-Language Models, Multimodal alignment, SFT, Preference optimization, Reward modeling, Agent development, Evaluation metrics (GSB, Rubric, Human-Model consistency), Data engineering, Quality governance
Preferred skills
Healthcare domain experience, Full-modal large model experience, Automated evaluation system construction, Large-scale online business experience
Technologies
PyTorch, SFT, Knowledge Distillation, Reward Model, Agent, GSB, Rubric
Responsibilities
Develop algorithms for cross-modal content understanding, entity extraction, classification, generation, and ranking; Build Judge Models and evaluation agents for Pointwise, Pairwise, and GSB testing; Construct gold standard, hard case, and adversarial datasets; Establish automated root cause analysis and data generation loops based on online bad cases; Collaborate with medical experts to deploy capabilities in health Q&A, dermatology, and medical imaging scenarios.
Seniority
Senior, hands-on IC