Rafiq | AI Automation Engineer

Business Operations Transformation

A full transition from fragmented, manual workflows into a scalable, automation-driven and AI-augmented infrastructure.

✖ Before

No automation. No database. No dashboards.

Operations relied on large spreadsheets that often crashed as data scaled
Data scattered across multiple spreadsheets, with no reliable historical records
Daily cross-team briefings just to report basic operational numbers
Order processing was fully manual, requiring large operational teams
Paid HR and attendance SaaS with low adoption and low compliance
No system-level integration between tools and workflows

✔ After

Automated workflows. Central database. Real-time dashboards.

Operations powered by databases and internal web apps instead of spreadsheets
Centralized data infrastructure as a single source of truth
Real-time BI dashboards replacing daily operational reporting
Automated order processing handling high-volume transactions with a lean team
Internal HRIS with AI-driven WhatsApp reporting, higher compliance, zero SaaS cost
Workflow automation established as a core business capability across all operations

Core Solution Pillars

Automation Core

Workflow Automation & Orchestration

Production-grade automation systems that replace repetitive, rule-based operations across teams, tools, and platforms.

API · Webhooks · Schedulers · Looping · Multi-Step Logic

AI Systems

AI Agent & Communication Systems

Intelligent chat and voice agents for routing, decision-making, alerts, and real-time interaction — internal or external.

WhatsApp · Telegram · Voice AI · Intelligent Routing · Alerts

Data Infrastructure

Unified Data & Analytics Infrastructure

Centralized data architecture that consolidates fragmented operational data into a single, reliable source of truth.

Databases · ETL · Dashboards · Reporting

Cost & Control

Self-Hosted Internal Tools (SaaS Replacement)

Custom-built internal tools deployed on your own infrastructure to replace recurring SaaS subscriptions without losing capability.

Sales · Marketing · CRM · Operations · HR · Finance

Selected Case Studies

🥇 Flagship · Operations Without Limits

E-Commerce Automation Engine

Python · Docker · Supabase · APIs · Control Panel

Problem: Managing high-volume orders across 4 stores required excessive admin resources and manual processing.
Solution: A multi-worker autonomous engine with a web-based control panel to manage tokens, monitor bot status, and delivery chat 24/7.
Impact: 70% reduction in manual operations, 3× increase in operational throughput.

20k+ Orders

70% Work Reduced

3× Productivity

Architecture Overview

A distributed, event-driven automation system designed to handle high-volume order processing across multiple stores.

Incoming order events are received via webhook push from marketplace platform
A centralized queue distributes tasks to multiple servicess
Each service processes orders independently (fetch → validate → fulfill → update)
Supabase acts as the single source of truth for order state and logs
A web-based control panel provides real-time monitoring, manual override, and process management
Retry and fallback mechanisms ensure reliability under high load

System Diagram

            [ Marketplace Platform ] 
                 (Webhook Push) 
                        │
                        ▼
               [ Webhook Service ] 
                        │
                        ▼
                    [ Redis ] 
              (Event Buffer & Queue) 
                        │
                        ▼
               [ Ingestor Service ]
              (Fetch Order via API)
                        │
                        ▼
                  [ Supabase DB ] 
               (Orders, State, Logs)
                        │ 
       ┌────────────────┼────────────────┐
       ▼                ▼                ▼
[ Fulfillment ]    [ Shipping ]    [ AI Chat Bot ]
    Service          Service          Service
       │                │                │
       ▼                ▼                ▼
[ Delivery API ]  [ API Update ]  [ Auto Reply ]
       │
       ▼
  [ Customer ]

Production Infrastructure

Designed not just as an automation system, but as a production-grade, self-hosted infrastructure capable of running continuously with minimal supervision.

Containerized microservices deployed via Docker across multiple VPS environments
Self-hosted infrastructure for full control, scalability, and cost efficiency
Integrated internal tools (e.g. voucher redemption system) connected to fulfillment workflows
Real-time monitoring and error tracking via Sentry
Uptime and service health monitoring using UptimeRobot
Secure tunneling and domain management via Cloudflare
Retry, logging, and failure handling across all services to ensure system reliability

🥈 Strategic · Zero-Friction Attendance

AI-Powered WhatsApp HR Operations (Agentic)

n8n · WhatsApp API · AI Agents · Internal Dashboards

Problem: Recurring HRIS costs and inefficient reporting through static workflows.
Solution: A stateful AI agent on WhatsApp that collects, validates, and structures employee reports through interactive conversations.
Impact: 99% reporting compliance, zero SaaS cost, higher data quality with minimal manual review.

99% Compliance

AI Agent Driven

$0 SaaS Cost

Architecture Overview

A hybrid system combining event-driven automation with a stateful AI interaction layer.

WhatsApp as the primary interface (via self-hosted WAHA gateway for full control and zero API cost)
n8n orchestrates workflows and triggers
AI agent handles context-aware interaction and validation
Supabase stores both conversation state and structured outputs
Dashboards display only validated, AI-generated summaries

System Flow

User → WhatsApp (WAHA) → n8n → AI Agent ↔ Supabase (State)
                         ↓
                 Validation Loop
                         ↓
                 Structured Summary
                         ↓
            HR Portal (Web Dashboard)
                         ↓
          Attendance System (MySQL Sync)

Agentic System Design

Stateful memory: tracks conversation context per user
Validation loop: AI asks follow-up questions until data meets standard
Structured output: only validated summaries are stored
Adaptive interaction: replaces static forms with dynamic questioning

AI Infrastructure

Powered by 9router for:

Multi-model routing & fallback
Cost optimization per request type
Scalable AI usage with predictable cost

System Integration

Connected to a custom HR web portal for monitoring and manual attendance control
Synchronizes with the attendance system via database-level state changes

Key Upgrade

Evolved from rule-based automation into a context-aware AI agent that evaluates and improves human input in real-time.

🥉 Operational · Single Source of Truth

Unified Data Bridge & Real-time BI

n8n · Supabase · Looker Studio · SQL · Spreadsheet

Problem: Cross-division reporting relied on manual aggregation, with hours wasted preparing basic metrics before meetings.
Solution: A centralized data infrastructure combining scheduled data pipelines and real-time operational data, powering unified BI dashboards across all divisions.
Impact: Eliminated manual daily reporting, 100% automated KPI tracking, and zero prep-time for leadership meetings.

5+ Automations

10+ Dashboards

Live BI Insights

Architecture Overview

A centralized data infrastructure that consolidates fragmented operational data into a dual-layer system for both real-time operations and scheduled analytics.

Data is extracted from multiple sources (spreadsheets, APIs, websites)
n8n orchestrates scheduled ETL pipelines for cleaning and transformation
Supabase serves as the unified data layer for orders, customers, and real-time inventory
Supabase views provide consistent, query-ready datasets
Looker Studio and internal web dashboards consume the same data layer
Combines real-time operational data (inventory) with scheduled analytical pipelines (orders, reporting)

System Diagram

 [ Spreadsheets ]   [ APIs ]   [ Websites ]
        │              │              │
        └──────────────┴──────────────┘
                       │
                       ▼
                  [ n8n ETL ]
             (Scheduled Pipelines)
                       │
                       ▼
                [ Supabase DB ]
               (Orders, Customers)
                       │
                       ▼
               [ Supabase Views ]
                       │
         ┌─────────────┴─────────────┐
         ▼                           ▼
 [ Looker Studio ]           [ Internal Web App ]
    (Analytics)                 (Admin Panel)

--------------------------------------------

            [ Stock Generator App ]
                       │
                       ▼
            [ Real-time Stock DB ]
                       │
                       ▼
             [ Internal Web App ]
            (Live Inventory View)

Data Architecture Layers

Operational Layer (Real-time):
Real-time inventory system used as the single source of truth for stock availability, directly consumed by internal web applications

Analytical Layer (Scheduled):
Order and customer data processed via scheduled pipelines for reporting, dashboards, and decision-making

Data Modeling & Structure

Unified 3 order sources (WhatsApp, marketplace, and web) into a single dataset
Designed a consistent relational schema for orders and customers
Created derived views for analytics and dashboard consumption
Linked transactions to unified customer profiles for cross-channel analysis

Multi-Source Data Integration

WhatsApp orders: Spreadsheet → n8n → Supabase
Marketplace orders: Webhook push → internal gateway → Supabase
Web orders: Direct insertion from internal web apps
All sources are normalized and merged into a unified order pipeline

Scheduled Data Pipeline

Automated ETL pipelines using n8n and Supabase functions on scheduled intervals
Designed for reliability over unnecessary real-time complexity
Ensures consistent data updates across all dashboards
Eliminates repetitive reporting workflows

Hybrid Workflow Design

Combines automation with manual spreadsheet workflows for multi-admin operations
Preserves existing processes to ensure adoption and usability
Supports gradual transition from manual reporting to automated systems
Balances system efficiency with real-world operational needs

Decision Layer

Data exposed via Supabase views for consistent querying
Integrated into Looker Studio and internal web dashboards
Embedded analytics within operational web apps for direct usage
Provides real-time BI experience on top of scheduled data pipelines
Eliminates daily reporting overhead with self-serve analytics

View Full Project Archive →

Crisis Management

Leadership Under Pressure

When a critical infrastructure failure threatened to halt entire business operations, I led the emergency recovery. Within 3 hours 24 minutes, I successfully restored core services, re-established secure API connections, and resumed automated workflows — preventing significant revenue loss.

3 Hours 24 Min Total Recovery Time

100% Core Operations Restored

R&D

From Idea to MVP

Exploring the next layer of autonomous systems: AI agents, voice interfaces, and digital twins.

🌍 WindSiter: Geospatial AI

🎭 Live2D AI Character

📞 Intelligent Voice Agent

🤖 AI-Powered Digital Twin

🌱 ESG Sustainability Platform

🏢 Consulting Marketplace Hub

MVP (Pre-Production)

Autonomous Voice Receptionist

Voice AI capable of holding natural phone conversations, routing calls, and handling inquiries without human intervention.

MVP (Pre-Production)

Context-Aware Internal VA

Multi-intent conversational assistant that routes employee requests across IT, HR, and reporting workflows. Identifies intent and dispatches to the appropriate sub-workflows.

Development

RAG-Powered IT Support Agent

Autonomous IT support agent with a self-managing RAG knowledge base. Auto-embeds and auto-purges vectors of documents on Drive triggers.

How I Work

System Design & Integration

I design end-to-end automation systems and integrate them into your existing tools without disrupting ongoing operations.

Build, Deploy & Own

I build and deploy production-ready systems using secure, self-hosted infrastructure so you fully own and control them.

Automate & Optimize

I continuously refine workflows to reduce manual work, improve reliability, and increase operational throughput.

Why Me

I build systems, not scripts

Moving beyond simple automation scripts to production-grade, observable infrastructure.

I replace operations, not assist them

Designing for full autonomy so you can scale without adding headcount.

I design for scale, not prototypes

Architecting solutions that handle enterprise-level volume and complexity.

The Engineer Mindset

With a background from Bandung Institute of Technology (ITB), Indonesia's top engineering university, I approach automation as system engineering — prioritizing reliability, observability, and long-term scalability.

I don't just build automations — I design systems that scale and survive without me.

20,000+ Transactions Automated
70% Manual Work Reduced
3× Operational Throughput

Ready to remove bottlenecks?

Available for remote contracts and system consulting.

Email Me LinkedIn