Agent Development & Automation
- Design, build, and maintain Claude-hosted agents that support Corkboard’s departments (paid media, SEO, web, creative, email/SMS, social, reporting)
- Write and deploy automations and scheduled workflows that handle routine tasks and free up team capacity
- Translate manual team processes into well-scoped, dependable, maintainable agent workflows
- Extend existing systems (such as the AVA assistant and its daily briefing routines) rather than rebuilding from scratch, always favoring the simplest working solution
Data Integration & API Management
- Write and maintain integrations against marketing platform APIs (Meta, Google, GA4, Search Console, TikTok, Amazon Ads, and others)
- Build and maintain the in-house data layer that pulls reporting data directly from the source, reducing reliance on costly third-party middleware
- Implement authentication flows, handle API errors and rate limits gracefully, and manage OAuth connectors and credentials securely (Slack, monday.com, Google Workspace, Gmail, etc.)
- Evaluate and integrate third-party tools (such as MCP servers) only when they solve a clear, confirmed need
Skills, Prompts & Tooling
- Build reusable “skills” and prompt templates that give agents consistent, on-brand, repeatable behavior
- Write clear, well-structured prompts and iterate based on real output quality
- Maintain the internal codebase and tooling with version control, sensible structure, and clear documentation
- Document tools so non-technical team members can use and trust them without deep technical knowledge
Quality Assurance & Human Oversight
- Uphold Corkboard’s human-in-the-loop model: agents draft and stage work for review, and no system posts or takes live action automatically
- Build in guardrails, review checkpoints, and clear status stages (In Progress → In Review → Approved)
- Test and debug agent output for accuracy, safety, and brand alignment before anything reaches a client
- Identify and resolve failures, edge cases, and unexpected behavior across the system
Performance Monitoring & Optimization
- Monitor agent and automation performance, logs, and daily outputs to catch issues early
- Track where systems save time or fall short, and refine based on real usage
- Improve the reliability, speed, and cost-efficiency of existing workflows over time
- Stay current on new AI models, tools, and best practices, and recommend where they’d add real value
Collaboration & Communication
- Work closely with department leads to understand their workflows and build tools that fit how they actually work
- Partner with the Co-Founder on the agency’s broader AI roadmap and priorities
- Explain technical concepts and tradeoffs in plain language to non-technical teammates
- Participate in planning discussions to align AI development with agency goals
