Senior Node.js Developer: Build AI Marketing Agent (Anthropic Claude, MCP, AWS, Slack)
I am a digital marketing consultant looking to build a custom AI media buyer and account manager for my agency. We currently manage 60 to 80 PPC clients with a team of three contractors.
The goal is to automate daily reporting, anomaly detection, and eventually routine campaign adjustments using Anthropic's Claude API and the Model Context Protocol (MCP). The AI will use Slack as its primary interface, allowing my team to interact with client ad accounts directly through chat.
What is already set up and provided to you:
- An AWS EC2 server (t3.small, Sydney region) ready to go.
- A Slack workspace with a custom App/Bot created and API tokens generated.
- Anthropic API keys.
- Master agency logins for Google Ads MCC, Meta Business Manager, and GA4.
- The "Brain": I will provide the exact system prompts, standard operating procedures, and strict logic rules based on 15 years of digital marketing experience.
What you will build:
- The Engine: A Node.js agent running 24/7 on the AWS server, managed securely by PM2.
- The AI Connection: Integration with the Anthropic API (specifically claude-sonnet-4-6).
- The MCP Bridge: Secure MCP connections to Google Ads and Meta Ads. You will need to build a robust system for managing OAuth tokens across dozens of client accounts without breaking rate limits.
- The Interface: Two-way Slack integration where the bot can drop scheduled reports, ping contractors with alerts, and read commands to fetch live data.
Project Rollout Phases:
- Phase 1 (Read-Only): Scheduled daily Slack drops pulling Spend, CPA, and ROAS from Google and Meta.
- Phase 2 (Alerts): Programmatic anomaly detection (e.g., sudden spend spikes or broken tracking) that pings the assigned contractor in Slack.
- Phase 3 (Execution): Enabling secure "write" access so the AI can execute basic ad platform changes based on explicit Slack commands from my team.
Requirements:
- Expert-level Node.js backend experience.
- Proven track record building complex API integrations and managing multi-client OAuth flows.
- Direct experience with Anthropic's API and MCP (Model Context Protocol) servers.
- Clear communication and the ability to hand over clean, documented code.
When applying, please start your proposal with the word "Performance" and include examples of any previous AI agents or complex API systems you have built.
Apply tot his job
Apply To this Job
The goal is to automate daily reporting, anomaly detection, and eventually routine campaign adjustments using Anthropic's Claude API and the Model Context Protocol (MCP). The AI will use Slack as its primary interface, allowing my team to interact with client ad accounts directly through chat.
What is already set up and provided to you:
- An AWS EC2 server (t3.small, Sydney region) ready to go.
- A Slack workspace with a custom App/Bot created and API tokens generated.
- Anthropic API keys.
- Master agency logins for Google Ads MCC, Meta Business Manager, and GA4.
- The "Brain": I will provide the exact system prompts, standard operating procedures, and strict logic rules based on 15 years of digital marketing experience.
What you will build:
- The Engine: A Node.js agent running 24/7 on the AWS server, managed securely by PM2.
- The AI Connection: Integration with the Anthropic API (specifically claude-sonnet-4-6).
- The MCP Bridge: Secure MCP connections to Google Ads and Meta Ads. You will need to build a robust system for managing OAuth tokens across dozens of client accounts without breaking rate limits.
- The Interface: Two-way Slack integration where the bot can drop scheduled reports, ping contractors with alerts, and read commands to fetch live data.
Project Rollout Phases:
- Phase 1 (Read-Only): Scheduled daily Slack drops pulling Spend, CPA, and ROAS from Google and Meta.
- Phase 2 (Alerts): Programmatic anomaly detection (e.g., sudden spend spikes or broken tracking) that pings the assigned contractor in Slack.
- Phase 3 (Execution): Enabling secure "write" access so the AI can execute basic ad platform changes based on explicit Slack commands from my team.
Requirements:
- Expert-level Node.js backend experience.
- Proven track record building complex API integrations and managing multi-client OAuth flows.
- Direct experience with Anthropic's API and MCP (Model Context Protocol) servers.
- Clear communication and the ability to hand over clean, documented code.
When applying, please start your proposal with the word "Performance" and include examples of any previous AI agents or complex API systems you have built.
Apply tot his job
Apply To this Job