大模型安全产品-豆包对话
Core
Establishing safety baselines for the Doubao conversational large language model through training, evaluation, and agent solution development.
Role type
Senior IC large model safety engineer (conversational LLM)
Builds
Production safety training pipelines, evaluation frameworks, and safety agent solutions for the Doubao product
Domain
Generative AI / Large Language Models / Conversational AI
Deliverable
production ML models
Required skills
Large model training (pre-training, fine-tuning, PE, RAG), Safety strategy operations, CV model experience, Evaluation framework design, Data curation and versioning, Experimental analysis
Preferred skills
Agent architecture design, User perspective analysis, Technical planning and drive
Technologies
Pre-trained language models, Fine-tuning, PE, RAG
Responsibilities
Define safety standards and generate training data for the model; Optimize evaluation processes including test set creation and quality improvement; Explore and build conversational safety agents
Seniority
Mid-level, hands-on IC
