About the Role
We are building an internal AI system that acts as a persistent strategic brain, content engine, and project coordinator.
This is not a chatbot role and not a prompt-writing role.
This is a systems architecture role.
Your job is to design and implement the infrastructure that allows multiple AI agents to:
-Remember past work and decisions
-Route tasks intelligently
-Improve over time through feedback
-Surface all activity in one clear control panel
If successful, this system will eliminate repetition, reduce cognitive overhead, and allow our work to compound instead of reset.
What You’ll Build
In this role, you’ll design and implement the core systems that allow our AI agents to function as a reliable, persistent, and improving system.
You will be responsible for:
1. Persistent Memory
Building a long-term, structured memory layer for agents
Creating a canonical knowledge base for rules, preferences, frameworks, and decisions
Implementing retrieval of relevant past context to inform new tasks
2. Agent Orchestration
Designing routing logic to assign requests to the appropriate agent
Assembling the right context from memory and the knowledge base for each task
Capturing and storing agent outputs with appropriate metadata
3. Central Dashboard (Notion-based initially)
Designing and implementing a dashboard that shows:
Requests, projects, outputs, feedback, and knowledge base entries
Creating clear views for:
“In Review”
“Active”
“Approved”
“Memory Updates”
Enabling simple actions such as creating requests, approving outputs, and triggering training updates
4. Feedback & Training Loop
Implementing an approve / revise / reject workflow
Automatically updating memory based on edits, feedback, and decisions
Ensuring traceability between requests, outputs, feedback, and memory updates
5. Reliability & Maintainability
Implementing logging, error handling, and retry mechanisms
Adding simple cost and usage controls
Producing clear documentation so the system can be maintained and extended
Responsibilities
Build persistent memory for multiple AI agents
Integrate agents with Notion and other tools
Design and implement a central dashboard
Create feedback loops so agents improve over time
Prevent context loss and duplication
Maintain documentation and system clarity
Optimize for reliability and simplicity
Preferred Stack
n8n or Make for orchestration
Notion API for dashboard and knowledge base
Supabase + pgvector or similar for vector memory
OpenAI / Anthropic APIs
Optional: Slack or email integration
Required Experience
Proven experience building end-to-end automation systems
Experience integrating AI APIs into real workflows
Strong systems thinking and architecture judgment
Experience with Notion, Airtable, or similar as structured systems
At least one implementation of retrieval-augmented generation (RAG) or vector memory
Clear documentation habits
What Success Looks Like
New request → agent output appears in dashboard automatically
Agents stop forgetting because memory is updated continuously
All work-in-progress and approved outputs are visible in one place
Weekly improvement in output quality due to feedback-driven learning
Engagement Structure
Initial contract: 6–10 weeks
15–25 hours per week
Goal: deliver a working Phase 1 system
Opportunity to extend or deepen engagement after Phase 1
Compensation
$30-60/hr depending on experience
Final rate depends on experience and demonstrated capability
Apply Now
Apply Now