Remote AI Video Production Architect

Datacontroller Toogeza Ltd.; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

toogeza is seeking an AI Video Agent Engineer to design multi-agent pipelines for an AI-driven video production system. You will architect agents for video analysis, narrative generation, editing orchestration, and output synthesis, and you will define prompts and rules for cross-agent collaboration.

You will build and maintain n8n-based workflows, optimize tool-calling and model selection, and champion modular, scalable architectures for multimodal content generation and memory-enabled

Qualifications

• Strong experience building multi-agent systems, including intent and sub-intent modeling.
• Expertise in agent orchestration and communication between components.
• Knowledge of summarization pipelines and context passing across agents.
• Hands-on coding in Python and JavaScript.

Responsibilities

• Design and implement a multi-agent video editing system with dedicated agents for analysis, narrative generation, and production.
• Define prompts, behavioral rules, and structured instructions for cross-agent interaction.
• Develop and maintain complex n8n workflows with dynamic routing and tool-calling logic.
• Integrate APIs for multimodal generation and data processing, ensuring data transformation across stages.
• Tune model interactions and optimize pipelines for quality, reliability, and efficiency.

Skills

Multi-agent systems

Intent modeling

Agent orchestration

Agent communication

Summarization pipelines

Context passing between agents

Python

JavaScript

Tools

n8n

n8n MCP

FFmpeg

LangGraph

LangChain

HuggingFace

Job description

toogeza is seeking an AI Video Agent Engineer to design multi-agent pipelines for an AI-driven video production system. You will architect agents for video analysis, narrative generation, editing orchestration, and output synthesis, and you will define prompts and rules for cross-agent collaboration.

You will build and maintain n8n-based workflows, optimize tool-calling and model selection, and champion modular, scalable architectures for multimodal content generation and memory-enabled