AI Agent Architect / Builder

Headquarters: Montreal, Canada
URL: https://www.contentjet.com


ABOUT CONTENTJET
ContentJet is a Canadian UGC creative agency that connects brands with creators to produce high-performing user-generated content at scale. We serve global clients across e-commerce, healthcare, and marketing verticals. Our core belief: every manual, repetitive task inside an agency can and should be replaced by an AI agent or automation. We use Claude Code CLI, Claude Cowork with MCP connectors, OpenClaw, n8n, VAPI, and custom multi-agent systems to eliminate manual work across every layer of our operations — from creator sourcing and outreach to content production, team coordination, and client delivery.
THE MISSION: REPLACE MANUAL WORK WITH AI AGENTS
ContentJet is on a mission to become a fully AI-native agency. That means every workflow that currently requires a human clicking, copying, pasting, searching, emailing, or following up manually should be handled by an AI agent instead. We’re not looking for someone to build chatbots — we’re looking for someone who can systematically identify manual bottlenecks across our operations and replace them with autonomous agent systems that run 24/7 without human intervention.
This includes: creator discovery and vetting, outreach campaign creation, brief reading and task assignment, content pipeline management, quality control, reporting, team notifications, client communication, and any repetitive operational task that currently eats human hours.
THE ROLE
You are the person who designs and builds complete AI agent systems from concept to production. You will own the architecture of multi-agent pipelines — defining agent roles, communication flows, tool integrations, and orchestration logic.
You will work extensively with Claude Code CLI to build, test, and iterate on agent systems directly from the terminal — writing production-grade code, managing project structures, and leveraging Claude’s coding capabilities for rapid development. You will use Claude Cowork with MCP connectors to orchestrate complex workflows that read from Monday.com, discover creators on TikTok/Modash, create Instantly campaigns, produce personalized documents, and coordinate across Slack channels — all without manual intervention.
You will deploy agents via OpenClaw, our open-source AI agent gateway, to make them accessible across multiple platforms (Slack, Telegram, web) with full DM and group chat support, tool management, and multi-model routing.
WHAT YOU WILL DO


Systematically audit ContentJet’s operations to identify every manual task that can be replaced by an AI agent or automation


Design and build multi-agent systems that fully replace manual workflows: creator sourcing, outreach, qualification, brief processing, campaign management, reporting, and team coordination


Use Claude Code CLI as your primary development environment to build, debug, and ship agent systems rapidly from the terminal


Build complex automation workflows using Claude Cowork with MCP connectors — orchestrating Monday.com, TikTok, Modash, Instantly, Google Drive, Slack, and other platforms into seamless, zero-touch pipelines


Deploy and configure agents via OpenClaw for multi-platform delivery (Slack, Telegram, web interfaces) with DM and group chat support


Architect agent communication flows, decision trees, tool-use schemas, and escalation logic within the OpenClaw gateway


Integrate agents with external tools via APIs and MCP connectors (Monday.com, Modash, TikTok, Instantly, Apollo, Apify, Firecrawl, ElevenLabs, Creatomate, etc.)


Build AI-powered creator sourcing and outreach pipelines — from Monday.com brief reading through creator discovery to Instantly campaign creation, with personalized Word document guides produced automatically for team members


Build AI voice agents using VAPI + Claude + Twilio/Telnyx for creator and lead qualification, replacing manual phone screening


Write production-grade Python agent code with proper error handling, logging, Slack integration (multi-channel), and monitoring


Design prompt systems at the systems level — agent instructions, OpenClaw tool configurations, Claude Cowork task definitions, output parsing, and quality control


Create CLAUDE.md project files for agent codebases following enterprise-grade standards (security, design philosophy, code organization)


Continuously look for new manual processes to automate — the goal is zero manual work for repetitive tasks


MUST-HAVE SKILLS


Python (production-level, not just scripting) — you write code that runs in production 24/7


Deep prompt engineering — system-level agent instructions, multi-step reasoning chains, tool-use orchestration, not simple chatbot prompts


Claude Code CLI proficiency — comfortable building, testing, and shipping entire projects from the terminal using Claude as a coding partner


Claude Cowork experience — ability to design and execute complex multi-step workflows using MCP connectors to orchestrate across platforms


Experience with OpenClaw or similar AI agent gateways/orchestration frameworks for multi-platform agent deployment


API integration experience (REST, webhooks, OAuth, API chaining) — you can wire any two systems together


Experience building multi-agent or multi-step AI systems (CrewAI, AutoGen, LangGraph, or custom builds)


Strong understanding of MCP (Model Context Protocol) — how to connect Claude to external tools, databases, and services


Understanding of agent design patterns: ReAct, tool-use, delegation, human-in-the-loop, autonomous decision-making


Experience deploying agents across multiple platforms (Slack bots, Telegram bots, web chat)


Strong English communication (written) — you’ll write agent specs, workflow documentation, and team guides


NICE-TO-HAVE
 


Familiarity with n8n (self-hosted) for workflow automation alongside agent systems


Background in UGC, creator economy, influencer marketing, or content operations


Experience with Monday.com API, Modash, TikTok Creator Marketplace, or similar creator tools


Experience configuring OpenClaw for multi-platform deployment with tool availability management and config troubleshooting


French or Arabic language skills (for MENA market projects)


WHAT SUCCESS LOOKS LIKE
Within 30 days: You’ve audited our current manual workflows, deployed your first multi-agent system via OpenClaw integrated with Slack, and shipped a working creator sourcing automation using Claude Cowork with MCP connectors.
Within 60 days: You’ve replaced at least 3 major manual workflows with autonomous agent systems. You’re independently designing agent architectures using Claude Code CLI and deploying them via OpenClaw with minimal direction.
Within 90 days: You’re leading AI agent delivery across 5–9 concurrent automation projects, have eliminated the majority of repetitive manual tasks in our operations, and are training team members on agent-powered workflows. Manual work at ContentJet is the exception, not the norm.
 
 
 

To apply: https://weworkremotely.com/remote-jobs/contentjet-inc-ai-agent-architect-builder