Job Description
1. Lead the core algorithm development for multimodal large models, design model structures and training tasks, solving key problems in multi-source data association understanding and structured representation.
2. Build pre-training and downstream adaptation frameworks based on global multimodal data (images, point clouds, text, etc.), enhancing the model’s generalization ability in intelligent analysis, business decision-making, and other scenarios.
3. Drive prototype implementation and performance improvement in scenarios like personalized solution planning, resource scheduling optimization, AIGC solution generation
Track cutting-edge research from CVPR, ICCV, NeurIPS, exploring the integration of embodied intelligence, edge-cloud collaborative reasoning, etc., leading to paper and patent outputs.
Requirements
1. PhD or excellent Master’s in computer science, AI, electronics, communications, or related fields.
2. Deep understanding of multimodal content comprehension and large model mechanisms, expertise in pre-training optimization, cross-modal alignment, and other core technologies.
3. Proficient in PyTorch/TensorFlow, experience in multimodal data processing and model engineering.
4. Strong engineering implementation skills (Python/C++, etc.), with top conference papers, open-source projects, or related internship experience preferred.
You Will Gain:
1. Contribute to core multimodal large model