Back to Projects
Platinum TierAI Agents
Personal AI Employee — Hybrid Cloud-Local Autonomous Agent
A 24/7 autonomous AI agent that monitors Gmail, WhatsApp, and filesystem — drafts responses on a cloud VM while you're offline, then executes with your approval when you reconnect.
Last updated March 2026
Overview
A hybrid cloud-local autonomous AI agent that evolved through four architectural tiers (Bronze → Silver → Gold → Platinum). The system monitors multiple sources (Gmail, WhatsApp, filesystem), processes tasks through a git-synced Obsidian vault pipeline, and executes actions via MCP servers with human-in-the-loop safety gates.
The cloud VM (Ubuntu 24.04) runs 24/7, detecting emails and drafting responses. When the local laptop reconnects, a 60-second git-sync cycle pulls pending actions for human approval. This hybrid architecture enables true offline tolerance — the agent works while you sleep.
The Four Tiers
| Tier | Name | What It Does |
|---|---|---|
| Bronze | File Watcher + Vault Processing | Filesystem monitoring via Watchdog, YAML-frontmatter metadata, risk-based routing |
| Silver | Multi-Source Orchestration | 4 watchers (filesystem, Gmail, WhatsApp, scheduled), centralized priority queue, PM2 management |
| Gold | MCP Servers + External Integrations | 4 FastMCP servers, circuit breaker pattern, correlation ID tracking, health monitoring |
| Platinum | Hybrid Cloud-Local | Cloud VM drafting + local execution, 60-second git-sync, offline tolerance, live verified demo |
4 MCP Servers
| Server | Purpose | HITL Level |
|---|---|---|
| fte-email | Draft, send, search emails via Gmail API | Routine / Sensitive |
| fte-social | Post to Facebook, Instagram, Twitter/X | Sensitive / Routine |
| fte-odoo | Odoo 19 ERP — invoices, payments, summaries | Routine / Critical |
| fte-documents | Generate reports and CEO briefings | Routine |
Key Engineering Patterns
- Circuit breaker pattern (3 failures → 300s cooldown)
- Correlation IDs:
corr-YYYYMMDD-HHMMSS-XXXX - Risk-based action routing (routine/sensitive/critical)
- Dry-run defaults for safety
- Atomic claim-by-move concurrency (prevents duplication)
- Stale detection (>48h pending, >7d rejected)
- Secrets isolation (3 layers: .gitignore, pre-commit hook, cloud audit)
- 21 Architecture Decision Records (ADRs)
Live Demo (Verified March 12, 2026)
Correlation ID: corr-2026-03-12-d16b3470
| Step | Event | Agent |
|---|---|---|
| 1 | Test email sent to Gmail | User |
| 2 | Gmail watcher detected (120s poll) | Cloud VM |
| 3 | Orchestrator drafted → Pending_Approval/ | Cloud VM |
| 4 | Git-sync pushed to remote | Cloud VM |
| 5 | Local git pull fetched file | Local |
| 6 | User moved to Approved/ | Local |
| 7 | Approval watcher sent via Gmail API | Local |
| 8 | File moved to Done/ (status: completed) | Local |
Interested in similar work?
I build AI agents, full-stack applications, and cloud-native systems. Let's discuss your next project.
Get in TouchOther Projects
Platinum TierSaaS Products
CareerCoach Pakistan
Full-stack SaaS — paste a job description, answer 10 AI-generated questions in English or Urdu, and get instant scored feedback. Subscription billing, Google auth, transactional email, and analytics included.
Next.js 16SupabaseStripeGroq AIResend
ML Pipelines
AI/ML Job Market Intelligence Pipeline
End-to-end ML pipeline that scrapes RemoteOK, clusters jobs with K-Means + TF-IDF, classifies seniority across 3 models, generates LLM market insights via Groq LLaMA 3.3 70B, and pushes automated reports to Google Sheets — all from one API call.
Pythonscikit-learnFastAPIPostgreSQLGroq API
Platinum TierAI Applications
BusBot Pakistan
Pakistan's first AI-powered public bus guide — speak in Urdu, get your route. Built for Google AI Seekho 2026 competition, covering Lahore, Karachi, and Islamabad with community-driven route data.
ReactGemini 3 FlashFirebase FirestoreWeb Speech APIGoogle Cloud Run