Edge通用智能模型研究员 - Seed Model
Core
Researching intrinsic reward mechanisms, long-term memory, multimodal perception, and tool-use/agent capabilities to advance general intelligence and self-updating models.
Role type
Research Scientist (General Intelligence / Seed Models)
Builds
Next-generation efficient reasoning models, long-sequence reasoning systems, and multimodal fusion tools.
Domain
Artificial Intelligence, General AI, Multimodal Learning, Robotics
Deliverable
research
Required skills
Intrinsic reward mechanisms, long-term memory modeling, multimodal perception, tool-use and agent modeling, scalable knowledge extraction from low-SNR data
Preferred skills
Foundation algorithms, machine learning theory, computer vision, AIGC, NLP, reinforcement learning, C/C++ or Python programming, top-tier conference publications (CVPR, NeurIPS, etc.), competitive programming awards
Technologies
C/C++, Python, MLLM, GenMedia, AI for Science
Responsibilities
Explore intrinsic reward mechanisms to enable active learning and self-updating; Build long-term memory mechanisms for efficient reasoning; Research multimodal perception boundaries and fusion tools; Develop tool-use and action capabilities for visual systems; Propose new research topics from the bottom up; Prepare and present a research proposal during the interview process.
