AI & Multi-Channel Marketing Automation Systems Specialist (n8n / Claude / LinkedIn API)

Project Overview We are looking for an experienced AI Automation Engineer to build end-to-end autonomous marketing and outreach agents. The system must automate content creation using Claude 3.5 Sonnet / Anthropic API, crawl targeted LinkedIn connection data (specifically law enforcement/public safety connections), auto-distribute content across multiple brands/entities, and run multi-channel outreach sequences via Email and Social Media. Detailed Scope of Work & Deliverables Module 1: Profile Connection Crawling & LinkedIn Auto-Sharing Target Identification & Enriched List: Build an agent that securely interfaces with our LinkedIn account to map law enforcement, emergency management, and public safety connections. Data Enrichment: Enrich target profile metadata (Name, Agency, Title, Email) using tools like Clay, Apollo, or PhantomBuster into a centralized CRM / database (Airtable / PostgreSQL). Content Broadcast Agent: Setup automated publishing workflows using Unipile API or official LinkedIn API to post brand updates directly to personal and company profile feeds. Module 2: AI Multi-Brand Content Generation Engine (Powered by Claude) Multi-Brand System Prompts: Build modular Claude (Anthropic API) prompt templates capable of generating tailored marketing material for multiple distinct business brands/verticals. Format Flexibility: Automatically format outputs into LinkedIn text posts, image carousel outlines, Facebook posts, and long-form email newsletters. Human-in-the-Loop Review Dashboard: Integrate a lightweight review interface (Slack / Telegram / Airtable) where generated posts can be approved or edited with one click before publishing. Module 3: Multi-Channel Outreach & Campaign Execution Cold Email Sequences: Build automated email outreach pipelines (Instantly.ai, Smartlead, or SendGrid) with custom variables generated by Claude. Direct Outreach Automation: Trigger personalized LinkedIn direct messaging or connection requests to identified law enforcement contacts, respecting daily safe rate limits. Unification & Tracking: Centralize incoming replies and interaction statuses into a central master tracking dashboard. Technical Stack & Architecture Preferences Workflow Orchestration: n8n (preferred for control/privacy) or Make.com. LLM Model: Anthropic Claude (Claude 3.5 Sonnet API). LinkedIn/Social API: Unipile API / Linked API (API-first to prevent account flags). Database / CRM: Airtable or PostgreSQL. Email Infrastructure: Instantly.ai / Smartlead / Mailgun. Anti-Ban & Compliance Constraints (Strict Requirement) All automated actions on LinkedIn must incorporate randomized delays (e.g., 45–120 seconds between operations) and strict daily volume caps (max 20–30 outreach actions/day) to keep accounts safe. Workflows must rely on session key/API-based connections (like Unipile) rather than aggressive client-side browser scraping.