CareerPlanSign in

Video Generation Content Understanding and Feedback Researcher

United Kindom-London💼 Full-time🗓 2026-07-13 → 2026-09-26

Core

Researching model capabilities to understand generated video content, including semantics, objects, relationships, and spatial structures, while implementing state tracking and consistency modeling for real-time interactive generation pipelines.

Role type

Researcher (Video Generation Content Understanding and Feedback)

Builds

Proprietary technical roadmaps and evaluation systems for controllable interactive video generation in game scenarios.

Domain

AI Research, Video Generation, Game AI

Deliverable

production ML models

Required skills

Video understanding, Video prediction, Reinforcement learning, Multimodal generation and understanding, Diffusion model principles, Model training pipeline design, Python, PyTorch

Preferred skills

Game AI, Simulation environments, Closed-loop systems, Publications in top-tier conferences

Technologies

PyTorch, Diffusion models

Responsibilities

Research model ability to parse semantics and spatial structures in generated content, Implement state tracking and consistency modeling, Explore unified generation-understanding architectures, Research causal consistency control and physical constraints, Collaborate with Agent/RL teams for end-to-end closed loops, Validate controllable interactive capabilities in game scenarios, Track industry work and formulate technical roadmaps

Seniority

Senior, hands-on IC

Sourced via workday · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.