具身多模态交互算法专家-火山引擎
Core
Design, develop, and optimize multimodal interaction algorithms (voice, vision, text, touch) for embodied intelligence scenarios, focusing on data fusion, semantic understanding, and intent recognition.
Role type
Senior IC embodied AI multimodal algorithm engineer
Builds
Embodied intelligence interaction systems and datasets
Domain
AI / Robotics / Consumer Electronics
Deliverable
production ML models
Required skills
Multimodal large language models, Deep learning frameworks (PyTorch/TensorFlow), RAG systems (vector/ES/graph search), Distributed training (LoRA/P-Tuning), Data engineering for multimodal data, Model deployment optimization
Preferred skills
Voice interaction, Computer vision, Personalized recommendation systems, Automotive/home/embodied scenario data expertise
Responsibilities
Design and optimize multimodal interaction algorithms; Build and curate domain-specific datasets; Collaborate with product teams to define interaction logic and technical metrics; Optimize RAG and model deployment for low latency; Write technical documentation.
Seniority
Senior, hands-on IC