Google Cloud set up

We are building a secure, cloud-based AI infrastructure for internal financial analysis. You will deploy a Vertex AI Workbench (Python-based) on Google Cloud, configure a modular file system, install essential packages, and ensure compatibility with our internal AI agent (via OpenAI API).
This is a pure environment deployment—no trading logic or modelling is involved.
________________________________________
? What You Will Do:
• Deploy Vertex AI Workbench (Python 3.11+, Jupyter Notebook)
• Install and verify required packages:
o pandas, numpy, requests, matplotlib, streamlit, openai, langchain, tiktoken, sqlite3
• Create project file structure:
• /app/
• ├── agents/
• ├── data/
• ├── output/
• ├── config/
• └── web/
• Enable and configure:
o Vertex AI API
o Cloud Scheduler (trigger every 5 minutes)
o Secret Manager (for OpenAI key – stored by us)
o Cloud Storage (simple test bucket setup)
o BigQuery (enabled only; no setup required)
• Install Streamlit and host a basic test dashboard
• Create and run a dummy script via Cloud Scheduler (e.g., log timestamp to /app/data/)
• Provide detailed documentation for:
o Access/reset instructions
o OpenAI API test confirmation
o Scheduler verification
o Cost management best practices
________________________________________
? Security Requirements (Strict):
• You will not access any proprietary, trading, or production code
• All code must be clean, human-readable, and unminified
• No encrypted logic, no obfuscation, and no compiled binaries
• You will not install any outbound traffic tools, webhooks, remote shells, or network tunnels (e.g., ngrok, socket, etc.)
• All files, scripts, and settings must be visible and auditable
• We reserve the right to audit all scripts line-by-line before activation
• Final access credentials and service accounts must be returned and rotated after delivery
________________________________________
? Stack:
• Google Cloud Platform (Vertex AI Workbench preferred)
• Python 3.11+
• Streamlit (for dashboards)
• Cloud Scheduler, Secret Manager, Cloud Storage
• OpenAI API integration (test key provided)
• Optional: BigQuery (just enable, no config)
________________________________________
? Deliverables Checklist:
• ✅ Fully deployed Vertex AI Workbench notebook
• ✅ Required Python packages installed and logged
• ✅ Verified folder structure under /app/
• ✅ Sample script that logs timestamp via Cloud Scheduler
• ✅ OpenAI API test call with successful log
• ✅ Basic Streamlit dashboard that reads a local log file
• ✅ All services (Scheduler, Storage, Secrets) properly enabled
• ✅ IAM roles and credentials created (with minimal permissions)
• ✅ Documentation covering setup, access, reset, cost tips
________________________________________
depending on:
• Speed of delivery
• Documentation clarity
• Code structure and modularity
Bonus available for outstanding delivery and clear commenting.
________________________________________