Spatial Reasoning Evaluation Specialist - Video RL

Role Overview:

We are looking for Spatial Reasoning Evaluation Specialists to support a Video Reinforcement Learning (Video RL) program focused on AI-generated videos, autonomous agents, physical-AI simulations, and spatial reasoning. The role involves reviewing AI-generated video outputs and evaluating whether the AI is performing actions correctly, logically, and realistically.

Key Responsibilities:

• Review AI-generated videos frame-by-frame and evaluate the correctness of actions and movements.
• Check temporal sequence, action order, and cause-effect relationships in videos.
• Evaluate whether AI-generated actions are aligned with the intended goal or task.
• Identify spatial, physical, visual, and motion-related errors in AI outputs.
• Review video segmentation, bounding boxes, masks, object boundaries, and avatar/character movements.
• Identify issues such as visual hallucinations, unnatural physics, clipping, incorrect movements, and inconsistencies.
• Provide clear, structured and evidence-based feedback explaining why an AI output is correct or incorrect.
• Support the creation and validation of ground-truth datasets and edge-case scenarios.

Mandatory Requirements:

Bachelor’s/Master’s degree in STEM - Computer Science, Engineering, Physics, Mathematics, Data Science or related field.

Relevant exposure to at least one of the following:

• Video Analysis
• Computer Vision
• AI/ML
• Reinforcement Learning
• Robotics / Simulation
• Video/Data Annotation
• AI Model Evaluation
• Strong spatial and temporal reasoning skills.
• Excellent attention to detail, particularly when analyzing visual content frame-by frame.
• Ability to identify logical, physical and visual inconsistencies in videos.