GenAI Researcher

Tencent•Palo Alto, CA

2d•Onsite

About The Position

This role is responsible for the long-term technical roadmap planning, core architecture design, and model research & development for Tencent Video's film and television AIGC business. The focus is on scenarios such as film & television video model generation and post-editing. The position involves leading the research and industrial application of cutting-edge generative models, building high-fidelity, controllable, and temporally consistent video generation and editing models. The technical lead will drive team innovation, define technical directions, manage project progress and quality, and enhance the team's overall technical capabilities. Continuous tracking of academic and industry advancements (e.g., Sora, Veo, Kling, Runway) is expected to maintain technical leadership in video generation.

Requirements

Master’s or Doctoral degree in Computer Science, Artificial Intelligence, Multimedia or other related majors.
Over 2 years of algorithm R&D experience, with the latest 3 years deeply focused on AIGC video generation and editing.
Proficient in Python and skilled in deep learning frameworks including PyTorch and TensorFlow, with strong engineering implementation and code optimization capabilities.
Solid theoretical foundation in machine learning and deep learning, with in-depth understanding and rich hands-on experience of generative models (Diffusion Model, Flow Matching, auto-regressive model, VAE, Transformer, etc.).
Familiar with large-scale distributed training technologies (e.g., DeepSpeed, FSDP, Megatron-LM).
Have representative achievements in video generation and editing, with relevant papers published in top conferences including CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML and SIGGRAPH.

Nice To Haves

Candidates with experience in training and optimization at the kilocard scale or above are preferred.

Responsibilities

Take charge of the long-term technical roadmap planning, core architecture design and model research & development for the film and television AIGC business of Tencent Video, covering scenarios such as film & television video model generation and post-editing.
Lead the research and industrial application of cutting-edge generative models (e.g., DiT, unified understanding-generation architecture). Build video generation and editing models with high fidelity, strong controllability and long-term temporal consistency.
Serve as the technical lead to drive continuous innovation of the team, define technical development directions, control project progress and delivery quality, and elevate the overall technical capability of the team.
Continuously track the latest advances in academia and industry (e.g., Sora, Veo, Kling, Runway). Evaluate and introduce cutting-edge technologies and methodologies to maintain the team’s technical leadership in the video generation field.