Research Scientist, Multimodal Foundation Models
Tencent AI Lab
- Location
- Hong Kong
- Level
- Research Scientist
- Posted
- 21 Aug 2026
ResearchResearch ScientistOn-siteActive
Original posting
Skills
- Multimodal
- Vision-language
- LLM
- PyTorch
- Evaluation
Summary
Tencent AI Lab's Hong Kong office is hiring a research scientist for multimodal foundation models. The project connects vision encoders with a language backbone used by WeChat and Tencent Video understanding features. You will propose training objectives for image-text and short-video data, run ablation studies, and write the evaluation that decides which checkpoint ships. The posting is on site in Hong Kong. Requirements: a research record in vision-language models or multimodal learning, solid PyTorch, and comfort presenting negative results as clearly as gains.
Full description
Tencent AI Lab's Hong Kong office is hiring a research scientist for multimodal foundation models. The project connects vision encoders with a language backbone used by WeChat and Tencent Video understanding features.
You will propose training objectives for image-text and short-video data, run ablation studies, and write the evaluation that decides which checkpoint ships. The posting is on site in Hong Kong.
Requirements: a research record in vision-language models or multimodal learning, solid PyTorch, and comfort presenting negative results as clearly as gains.