Join MVP Lab

Focus

• Research on ViT and MLLM models.
• Reasoning multimodal models.
• Large scale model training.

Requirements

• Strong coding ability and practical engineering implementation skills.
• Long-term commitment to AGI and active exploration of frontier multimodal technologies.
• Large scale ViT, LLM, VLM, or MLLM training experience is a plus.

Focus

• Build scalable training infrastructure for multimodal models.
• Optimize GPU utilization, cluster scheduling, large scale cluster performance, and end-to-end training throughput across large-scale MLLM workloads.
• Develop reliable model engineering systems for evaluation, inference, deployment, and fast iteration from research prototypes to production services.

Requirements

• Strong coding ability, systems design, and reliable implementation.
• Deep understanding of large-model frameworks such as PyTorch and distributed training stacks.
• Experience optimizing model architecture, training pipelines, inference performance, or infrastructure is strongly preferred.
• Large-scale training, GPU systems, cluster scheduling, or online model serving experience is a plus.