多模态算法工程师(J105498)
Core
Develop multimodal large models and intelligent problem-solving algorithms for the education sector, including layout analysis, OCR, content understanding, grading, and search relevance modeling.
Role type
Senior IC multimodal algorithm engineer (education)
Builds
Online problem-solving and grading Agents, offline content production systems, and self-evolving question banks.
Domain
Education technology + Multimodal AI
Deliverable
production ML models
Required skills
Multimodal understanding, Large model post-training, Search ranking, RAG, Agent design, Reward modeling, Automated evaluation
Preferred skills
Multimodal search, Photo-based problem solving, Education large models, Large-scale online system construction
Technologies
LLM-as-Judge, Task planning, Intelligent routing, Tool calling, Multi-turn interaction
Responsibilities
Design and build online problem-solving and grading Agents with core capabilities like task planning and RAG; Establish an offline content production and question bank self-evolution system; Build model capability evaluation and automated testing frameworks; Optimize model performance, latency, inference cost, and system stability for large-scale online business; Track and drive innovation in multimodal models and Agent technologies.
