AI Architect

VDart Inc; Los Angeles (CA); On-site; USD 170,000 - 210,000; Full time

Job summary

VDart Inc. in Los Angeles, CA is seeking an AI Architect to lead the design and delivery of agentic AI driven workflows. You will map current AI agent ecosystems, unify disparate agents into a single orchestration layer, and define a durable state model and runbook for production readiness.

You will collaborate with creative teams, integrate Google Workspace and design tools via APIs, and shape a pragmatic long-term AI strategy with measurable quick wins and reliable governance.

Qualifications

• 5+ years building and deploying production AI agent systems.
• Portfolio of deployed multi-agent orchestration work with architecture and reliability details.
• Experience with production-grade OpenAI, Claude, Gemini or equivalents.

Responsibilities

• Lead onsite discovery with a Visual Production team to map current workflows and bottlenecks.
• Unify independently built agents and automations across GPT, Claude, and Gemini into one orchestration layer.
• Deliver quick wins into daily production and shape a longer-term agentic AI strategy and roadmap.
• Design a canonical work-order schema and durable state model (job ID, assets, version lineage, routing, ownership, approvals, audit history).
• Develop conditional routing and human-in-the-loop approvals, including manual completion events for studio shoots and proofs.
• Create dynamic notifications and end-to-end visibility so the team knows when work is ready, stalled, or blocked.

Skills

Agentic AI

Workflow orchestration

Automation

Integrations

Python

TypeScript

REST APIs

OAuth

JSON schemas

Structured outputs

Tools

OpenAI

Anthropic Claude

Google Gemini

n8n

Make

Zapier

Google Apps Script

Figma API

Google Workspace API

Box API

MCP

Job description

Job Titile: AI Architect
Location: Los Angels CA
Contract

Projects:

• Leading onsite discovery with a Visual Production team of 6 8 to map current workflows, siloed AI agents, handoffs, approvals, and bottlenecks
• Unifying independently built agents and automations across GPT, Claude, and Gemini into one orchestration layer without rebuilding what already works
• Shipping quick wins into daily production use early, then shaping a longer-term agentic AI strategy and roadmap
• Designing a canonical work-order schema and durable state model covering job ID, assets, version lineage, routing, ownership, approvals, and audit history
• Building conditional routing and human-in-the-loop approvals, including manual completion events for studio shoots and print proofs
• Creating dynamic notifications and end-to-end visibility so the team knows when work is ready, stalled, or blocked
• Integrating Figma, Google Workspace (Drive, Slides), Box, and shared file repositories through APIs, webhooks, and MCP
• Propagating upstream changes safely to affected downstream work
• Documenting the system, training the team, and handing off a runbook and prioritized backlog

Experience

• 5+ years of hands-on experience building and deploying production systems in agentic AI, workflow orchestration, automation, or integrations
• Portfolio of deployed multi-agent or orchestration work solving similar problems, with the ability to walk through its architecture and reliability model
• Background as an Agentic AI Workflow Architect, Creative Technologist, AI Workflow Engineer, Solutions Architect, or Automation/Integration Architect
• Production agent development with OpenAI, Anthropic Claude, Google Gemini, or comparable platforms
• Workflow and state-machine design, including shared state, events, retries, idempotency, and conditional routing
• Python or TypeScript development with REST APIs, webhooks, OAuth, JSON schemas, and structured outputs
• Agent interoperability through MCP or equivalent approaches, with A2A, LangGraph, or Google ADK a plus
• Integration experience with Google Workspace APIs, Box, and DAM or shared file environments, with Figma API experience preferred
• Model-agnostic, vendor-neutral architecture built on stable interfaces
• Enterprise security fundamentals, including least-privilege access, credential management, and handling confidential assets
• Consultative and adaptable, comfortable in ambiguity and unorthodox environments, and able to fit best practices to how a creative team actually works
• Familiarity with creative operations such as image generation, retouching, or 3D is helpful, though technical depth comes first
• n8n, Make, Zapier, or Google Apps Script experience a plus