I build practical AI-powered automation workflows and AI agent solutions that connect AI models, business tools, APIs, and data to reduce repetitive work and improve productivity.
I am currently focused on strengthening my practical skills through hands-on automation projects while preparing to work on real-world AI automation solutions.
- 🤖 Focused on AI Automation, AI Agents, and Business Process Automation
- ⚙️ Building workflows with n8n, Make.com, and Zapier
- 🧠 Working with OpenAI, Claude, Gemini, RAG, and MCP
- 🔗 Integrating applications using REST APIs, Webhooks, JSON, and OAuth
- 📚 Continuously learning, building, testing, and improving practical automation solutions
- 🌍 Open to remote, hybrid, local, and international opportunities
- 📍 Based in Dhaka, Bangladesh
n8n • Make.com • Zapier
AI Agents • OpenAI • Claude • Gemini • RAG • MCP • Retell AI / Voice AI
REST APIs • Webhooks • JSON • OAuth / API Authentication • Web Scraping
Google Sheets • Gmail • Google Drive • CRM Automation • Lead Generation • Customer Support Automation
Supabase • Airtable • Apify • ElevenLabs • FFmpeg
Currently, I am focused on:
- Building end-to-end AI automation workflows
- Developing practical AI agent solutions
- Connecting AI models with external tools and business systems
- Improving API, webhook, RAG, and MCP-based integrations
- Documenting and publishing my automation projects on GitHub
An AI voice receptionist prototype developed as a hands-on automation project using Retell AI, MCP, and n8n.
The system connects a voice agent with multiple automation tools to handle course-related queries and business workflows.
Key Capabilities:
- Course search and class schedule checking
- Lead capture
- Consultation booking with Google Calendar
- Email confirmation
- FAQ knowledge-base support
- Human escalation request handling
- Conversation logging
Tech Stack:
Retell AI • MCP • n8n • Google Sheets • Google Calendar • Gmail • REST APIs • Webhooks
Project Highlights:
- 9 documented n8n workflows
- Multiple MCP-connected automation tools
- Voice AI + workflow automation integration
- Architecture diagram and workflow screenshots
- Sanitized public workflow files
- Documented limitations and future improvements
This is a learning and portfolio project developed as part of an AI Automation assignment and is not an official Ostad Academy production system.
An AI-powered news research automation developed as a hands-on project using n8n, GNews API, OpenRouter, and Google Sheets.
The workflow accepts a topic and article count through a webhook, researches and analyzes current news articles with an LLM, stores structured results, and returns the Top 3 stories ranked for LinkedIn relevance.
Key Capabilities:
- Topic-based news collection through GNews API
- Per-article AI summarization and research
- Practical key takeaway extraction
- AI news category classification
- LinkedIn Worthiness Score from 1–10
- Structured AI output validation
- Google Sheets research storage
- Top 3 article ranking by score and recency
- JSON API response through webhook
Tech Stack:
n8n • GNews API • OpenRouter • Google Sheets • Webhooks • JavaScript • Bruno
Project Highlights:
- End-to-end webhook-driven research workflow
- Configurable LLM through OpenRouter
- Structured Output Parser for consistent AI responses
- Custom JavaScript ranking logic
- Architecture diagram and workflow screenshots
- Sanitized public n8n workflow
- External workflow execution demo
- Documented setup, limitations, and future improvements
An end-to-end assignment collection and validation automation built with Make.com, Google Forms, Google Sheets, Google Drive, Google Docs, and Gmail.
The system processes student assignment submissions, cleans and validates submission data, detects duplicates, routes valid and invalid submissions, organizes records and folders, sends automated notifications, and updates final submission status in Google Sheets.
Key Capabilities:
- Google Forms-based assignment submission intake
- Google Sheets-triggered Make.com automation
- Data cleaning and normalization
- Duplicate submission detection
- Valid, invalid, and duplicate submission routing
- Student-wise Google Drive folder creation
- Automatic Google Docs submission summary generation
- Student confirmation emails
- Instructor notification emails
- Submission status tracking in Google Sheets
- On-time / late status tracking from submission data
Tech Stack:
Make.com • Google Forms • Google Sheets • Google Drive • Google Docs • Gmail
Project Highlights:
- End-to-end automated assignment processing workflow
- Three-route validation and submission handling
- Automated folder and document creation
- Student and instructor email notifications
- Centralized submission tracking in Google Sheets
- Architecture diagram and implementation documentation
- Sanitized public Make.com blueprint
- External workflow execution demo
- Documented test results and current implementation limitations
An end-to-end AI-powered real estate automation system built with n8n, Apify, Airtable, OpenAI, and MCP.
The system automatically collects Dubai rental property listings from PropertyFinder through Apify, cleans and structures the scraped data with AI, stores the listings in Airtable, exposes the database through an MCP Server, and allows an AI assistant to answer natural-language questions about available properties.
Key Capabilities:
- Scheduled Dubai real estate listing collection
- PropertyFinder scraping through Apify
- AI-powered property data cleaning and structuring
- Structured Output Parser validation
- Airtable listing storage and upsert by Listing URL
- MCP Server with 7 specialized property tools
- Budget-based property search
- Location-based listing search
- Latest listing retrieval
- Cheapest property lookup
- Listing count queries
- Full property list retrieval
- Natural-language property queries through an AI Assistant
Tech Stack:
n8n • Apify • Airtable • OpenAI • MCP • HTTP Request
Project Highlights:
- Three connected n8n workflows
- Automated scraping and structured property storage
- Seven Airtable-backed MCP tools
- AI Assistant connected through MCP Client
- Natural-language budget and location filtering
- Real chat-query examples with actual listing results
- Architecture diagram and workflow screenshots
- Sanitized public n8n workflow exports
- External demo video and Airtable database
- Documented setup, implementation notes, limitations, and cost estimation
An AI-powered email customer support automation built with n8n, Gmail, Supabase Vector Store, Google Drive, and RAG.
The system builds a searchable product knowledge base from a PDF, retrieves relevant information through a Supabase Vector Store, drafts context-aware customer support replies with an AI Agent, sends the generated reply for human approval, and delivers the final response to the customer only after approval.
Key Capabilities:
- Product knowledge-base ingestion from Google Drive
- Document loading and recursive text chunking
- Embedding generation for RAG
- Supabase Vector Store knowledge storage
- Gmail-based incoming customer email detection
- Full email message retrieval
- AI Agent response generation using retrieved knowledge
- Human-in-the-loop approval email
- Workflow pause and resume using Wait
- Approval validation through an If node
- Final customer email delivery after approval
- False branch intentionally left without a rejection-email action
Tech Stack:
n8n • Gmail • Google Drive • Supabase Vector Store • RAG • AI Agent • Embeddings
Project Highlights:
- Two-part RAG + customer support automation architecture
- PDF-based product knowledge ingestion
- Vector search through Supabase
- Retrieval-grounded AI email drafting
- Human approval before customer delivery
- Premium architecture diagram and workflow screenshots
- Sanitized public n8n workflow export
- External workflow execution demo
- Documented setup, implementation notes, and current limitations
- Model-flexible architecture without locking the workflow to a specific chat model
An AI-powered Facebook engagement automation built with n8n, Meta Webhooks, Google Sheets, AI intent classification, and Facebook Graph API.
The workflow receives Facebook Page comments through a webhook, identifies the related campaign and resource from Google Sheets, analyzes whether the commenter is requesting the resource, and routes the interaction through True/False branches.
Key Capabilities:
- Facebook Page comment capture through Meta Webhooks
- Campaign and resource lookup from Google Sheets
- Comment data normalization
- AI-based resource-intent classification
- Structured True/False routing
- Public Facebook comment reply for qualified requests
- Private Messenger DM with the requested resource link
- Dynamic resource delivery based on campaign data
- Positive-path validation with real test results
- Negative-path validation without unwanted messaging
Branch Logic:
- True branch: sends a public comment reply and delivers the resource link privately through Messenger DM
- False branch: sends no public reply and no private resource DM
Tech Stack:
n8n • Meta Webhooks • Facebook Graph API • Google Sheets • AI Intent Classification • Messenger Automation
Project Highlights:
- End-to-end Facebook Comment-to-DM automation
- AI-powered qualification before resource delivery
- Google Sheets-driven campaign/resource configuration
- True and False branch workflow evidence
- Positive and negative test screenshots
- Public reply + private DM delivery evidence
- Architecture diagram and workflow screenshots
- Sanitized public n8n workflow export
- External demo video, campaign post, Google Sheet, and project folder
- Documented setup, implementation logic, and test results
- 🌐 Portfolio: automation-ai-jamil.lovable.app
- 💼 LinkedIn: S. M. Jamilur Reza Razib
- 📧 Email: jamilur0101@gmail.com