Controllable Video Generation Research Intern

TencentSydneyHybridFull-time1 week ago

LIGHTSPEED STUDIOS is seeking a Controllable Video Generation Research Intern to design multimodal controllable generation frameworks for video, focusing on high-fidelity generation under sparse control signals and decoupling of control modalities. The role requires a Ph.D. in AI-related fields, proficiency in Python/PyTorch, and experience with conditional generation models and diffusion models.

What you'll do

Design a multimodal controllable generation framework supporting various inputs such as keyboard, controller, trajectory, and voice.
Research high-fidelity generation under sparse control signals and explore the extension of conditional control paradigms in the video domain.
Achieve the decoupling and combination of trajectory control, camera control, motion control, and posture control.
Design dedicated control modules tailored to different game genres.
Develop video generation capabilities for multi-player collaborative control, ensuring low-latency response and inter-frame state retention.
Collaborate with the Engine/Agent teams to define standardized interfaces for controllable signals.
Drive the join optimization of controllable generation and Reinforcement Learning post-training.

What you need

Must-have

PythonPyTorchComputer VisionDeep LearningConditional Generation ModelsDiffusion ModelsCamera ControlTrajectory Control

Nice-to-have

Game AIGame DevelopmentReal-time Video GenerationInteractive Video Generation

Soft skills

Collaboration

Video GenerationResearch InternComputer VisionDeep LearningGame DevelopmentAIPyTorchDiffusion ModelsHybridSydney

/What you'll do

/What you need

Must-have

Nice-to-have

Soft skills

What you'll do

What you need