Build AI agent Co-Worker
UpworkGBNot specifiedintermediateScore: 76
AI App DevelopmentAI ChatbotLLaMARetrieval Augmented GenerationAI Agent DevelopmentChatGPTLangChainAI BotTensorFlowPyTorchAI Model TrainingMachine LearningLarge Language ModelConversational AIGenerative AI
I am a Solutions Architect specialising in Anaplan (enterprise planning platform), and I am building a Proof of Concept for an AI agent “co-worker” that can assist with structured browser-based modelling tasks.
This is not a generic chatbot setup.
I am looking for someone who understands:
• AI agents (OpenClaw / similar agent frameworks)
• Browser automation (Playwright / Selenium / Chrome DevTools)
• Local-first RAG setups
• Secure sandboxed environments (macOS)
• Credential isolation & 2FA handling
• Prompt engineering for structured execution
The goal is to help me design and implement a controlled AI automation environment on a dedicated MacBook Pro.
⸻
🎯 Objective
Build a secure, isolated AI agent environment that can:
1. Control Chrome browser sessions
2. Navigate web applications (specifically Anaplan)
3. Execute structured UI tasks
4. Use a local knowledge base (Anapedia + internal standards)
5. Operate safely without exposing sensitive credentials
This is a Proof of Concept, not a production rollout.
⸻
🧠 Scope of Work
Phase 1 – Environment Setup
• Configure isolated macOS user environment
• Install and configure OpenClaw (or similar)
• Configure secure Chrome automation
• Implement session-based login strategy (manual 2FA, no password leakage)
• Restrict system permissions
Phase 2 – Knowledge Layer
• Set up local RAG pipeline
• Index curated documentation
• Configure retrieval and response validation
Phase 3 – Automation Testing
• Validate browser navigation reliability
• Test UI interaction with dynamic web app
• Debug timing / DOM interaction issues
• Ensure repeatability and stability
⸻
🔐 Security Requirements (Non-Negotiable)
• No storage of plaintext credentials
• No unvetted third-party skills/plugins
• Least-privilege system access
• Clear documentation of setup
• Reproducible environment
If you do not have experience building secure automation systems, this is not the right project.
⸻
✅ Ideal Candidate
• Experience with OpenClaw or similar AI agent frameworks
• Strong experience with Playwright or Selenium
• macOS automation experience
• Experience building local RAG systems
• Security-first mindset
• Comfortable working in iterative POC cycles
Bonus:
• Experience with enterprise SaaS platforms
• Experience automating complex React-based web UIs
⸻
📦 Deliverables
• Fully configured local agent environment
• Browser automation working reliably
• Local knowledge base integrated
• Clear documentation
• POC test report with findings
⸻
💬 To Apply, Please Answer:
1. What AI agent frameworks have you worked with?
2. Have you built browser-based automation for complex web apps?
3. How would you securely handle 2FA in an automation context?
4. How would you design a local RAG system for this use case?
5. Share links to similar projects (if possible).
⸻
🔍 Important
This is a strategic long-term direction. If this POC succeeds, there may be further work scaling the solution.
I am not looking for experimentation without structure. I am looking for a disciplined technical partner who can build this correctly.
Unlock AI Intelligence, score breakdowns, and real-time alerts
Upgrade to Pro — $29.99/mo