Hello, I’m

Nguyễn Quốc Khang.

DevOps Engineer
& AI Automation Builder

I build reliable infrastructure, connect business systems and develop AI assistants that help teams work with clearer information and fewer manual steps.

Ho Chi Minh City · Remote

Available for selected projects

Nguyễn Quốc Khang in a navy suit
NGUYỄN QUỐC KHANGDEVOPS & AI AUTOMATION
≈60Odoo / LXC production
workloads operated
11+Production n8n
automation workflows
9Department AI assistants
in one workspace
BUSINESS CAPABILITIES

Engineering for everyday business.

Reliable operations.

Cloud infrastructure, repeatable deployments, monitoring and clear recovery paths that your team can operate.

Explore the recovery case

Connected workflows.

Connect business tools with n8n, APIs and document workflows, with clear handoffs between people and systems.

See workflow examples

Accessible business knowledge.

AI assistants that retrieve relevant documents, show sources and work within defined access boundaries.

Explore GLX-AI · Active project
01 /SELECTED WORKReal systems. Real responsibilities.

Built to solve.
Designed to last.

A closer look at the infrastructure, automation and AI systems I build and operate.

CURRENT PROJECTS

What I’m building now.

Sales knowledge and business systems, connected with AI and clear access boundaries.

THE COMPLETE COLLECTION

Explore all 16 case studies.

Explore the problem, my contribution and the approach behind each implementation, operations case or architecture study.

16 / 16 CASE STUDIES
CLOUD ENGINEERING STUDIES

AWS media delivery & CI/CD.

Two additional studies in cloud delivery and repeatable infrastructure, from content distribution to automated deployments.

01Media Streaming CDN Exploration in AWS

Developed a website for formatted video delivery using AWS S3, CloudFront, MediaConvert, Lambda functions and Route 53 for reliable content delivery.

AWS S3CloudFrontMediaConvertLambdaRoute 53
02Terraform-based Jenkins CI/CD Implementation

Transformed website deployment into Terraform infrastructure as code with Kubernetes concepts on AWS and CI/CD pipelines powered by Jenkins.

TerraformKubernetesAWSJenkinsCI/CD
AUTOMATION, STEP BY STEP

11 workflows. See how they work.

From voice and SMS assistants to document, media and operations workflows. Open an example to explore its five steps.

01Backup to GitHub

Automated scheduled backup of n8n workflows to GitHub for version control, data safety, and operational history.

  1. Schedule Trigger

    Runs the backup process on a defined interval.

  2. Retrieve Workflows

    Fetches active workflows from a self-hosted n8n instance through its API.

  3. Format to JSON

    Normalizes workflow objects into versionable JSON files.

  4. Base64 Encoding

    Encodes file content for repository API updates.

  5. Sync to Repo

    Checks existing files and commits updates only when workflow content changes.

n8nVersion ControlBackupGitHub
02AI Chatbot for Facebook & Zalo-style Messaging

An AI chatbot workflow using business knowledge retrieval, Qdrant, RAG, and messaging webhooks for faster customer replies.

  1. Multi-Platform Webhook

    Receives incoming messages from supported messaging channels.

  2. Knowledge Ingestion

    Stores business documents, FAQs, and product data in a vector database.

  3. Semantic Search

    Retrieves the most relevant context from Qdrant before generating an answer.

  4. AI Response

    Uses an LLM to draft a context-aware response in the business tone.

  5. Escalation

    Routes low-confidence or sensitive conversations to a human team member.

n8nAI ChatbotQdrantRAGWebhooks
03Google Review Monitor & Auto-Reply

A review monitoring workflow that classifies sentiment, drafts AI replies, logs outcomes, and alerts the team.

  1. Review Monitoring

    Polls or receives new review data from supported APIs.

  2. Sentiment Analysis

    Classifies reviews as positive, neutral, or negative.

  3. Reply Drafting

    Generates brand-safe replies for team approval or publishing.

  4. Logging

    Stores review data, draft reply, and status in a spreadsheet.

  5. Alerts

    Notifies the team when urgent or negative reviews need attention.

n8nGoogle APIsSentimentReviewsAI
04SMS AI Center

A Twilio and n8n workflow for AI-assisted SMS conversations, customer inquiry handling, and team notifications.

  1. Twilio Webhook

    Receives incoming SMS messages in real time.

  2. Intent Classification

    Classifies the customer request and checks required context.

  3. Context Retrieval

    Fetches relevant business data or previous conversation notes.

  4. AI Reply

    Generates a safe response aligned with the business tone.

  5. SMS Delivery

    Sends the response and logs the interaction.

n8nTwilioSMSAI Agent
05Voice AI Call Center Workflow

A Vapi, LLM, and n8n orchestration workflow for voice AI calls, appointment flows, CRM lookups, and call logging.

  1. Voice Trigger

    Handles inbound or outbound voice events through Vapi.

  2. Conversation Engine

    Uses an LLM to understand caller intent and respond naturally.

  3. Tool Orchestration

    Lets n8n perform CRM lookup, booking, or data retrieval during the call.

  4. Call Flow

    Manages transfer, callback, voicemail, or escalation logic.

  5. Analytics

    Stores transcripts and call metadata for operational review.

VapiOpenAIn8nVoice AI
06Scrape Website to Notion or Sheets

A data extraction workflow for competitor research, lead tracking, and structured business data collection.

  1. Fetch Strategy

    Runs an HTTP or browser-based scraper depending on the target site.

  2. Extraction

    Uses AI or structured parsers to identify relevant business information.

  3. Cleaning

    Normalizes names, emails, addresses, and links.

  4. Database Mapping

    Writes clean data to Notion, Google Sheets, or another target.

  5. Alerting

    Sends a short summary when the workflow completes.

n8nWeb ScrapingAI AgentNotionGoogle Sheets
07Affiliate Video Generation Workflow

A marketing workflow concept that reads product links, extracts product context, creates prompts, and routes video assets for review.

Workflow concept
  1. Input Queue

    Reads product links and campaign metadata from a spreadsheet.

  2. Product Extraction

    Extracts product copy, image references, and selling points.

  3. Prompt Builder

    Creates a structured prompt for video or creative generation.

  4. Asset Routing

    Stores generated assets and captions for review.

  5. Publishing Prep

    Prepares captions and metadata for social platforms.

n8nAI VideoAffiliateMarketing
08TikTok/Douyin Dubbing & Subtitle Translation

A media automation workflow for downloading source media, transcribing audio, translating subtitles, generating voiceover, and rendering with FFmpeg.

  1. Content Intake

    Receives a video URL or file through a controlled input channel.

  2. Video Processing

    Normalizes video format and extracts audio with FFmpeg.

  3. Transcription

    Transcribes speech and creates timestamped subtitle data.

  4. Translation

    Uses an LLM to translate and preserve subtitle timing.

  5. Render

    Generates voiceover and renders subtitles into the final video.

n8nFFmpegGroq WhisperSubtitlesTikTok
09Facebook Page Comment Management Bot

A moderation and response workflow that classifies comments, drafts replies, tracks repeated issues, and alerts a human reviewer.

  1. Post Retrieval

    Fetches recent posts and comments from the connected page.

  2. Comment Processing

    Filters already-handled comments and extracts relevant context.

  3. AI Classification

    Classifies comment type, tone, and action priority.

  4. Reply Drafting

    Creates a suggested response for positive or support comments.

  5. Review Queue

    Logs risky comments for team approval before action.

n8nFacebookOpenAIModerationGoogle Sheets
10MCP Google Maps Workflow

A Model Context Protocol integration concept that exposes map search and place details to AI agents and n8n workflows.

Workflow concept
  1. MCP Server

    Exposes location tools to an AI agent in a controlled way.

  2. Place Search

    Searches for places or businesses from natural language requests.

  3. Details Extraction

    Retrieves structured place information and operational metadata.

  4. Pipeline Output

    Writes clean results into Sheets, a database, or downstream workflows.

  5. Agent Actions

    Lets the agent use the tool output to continue the business process.

n8nMCPGoogle MapsAI Agent
11Workflow Automation for Email Validation

A lead hygiene workflow that validates email format, checks DNS/MX data, scores addresses, and logs results for marketing teams.

  1. Webhook or Sheet Trigger

    Receives a list of email addresses from a form, sheet, or webhook.

  2. Syntax Check

    Filters malformed addresses before slower validation steps.

  3. DNS Analysis

    Checks domain and MX records for deliverability signals.

  4. Scoring

    Assigns a confidence score and reason for each address.

  5. Reporting

    Updates the source sheet or sends a summary to the team.

n8nEmail ValidationDNSData Quality
Explore my n8n workflow collection
02 /WHAT I BRINGFROM INFRASTRUCTURE TO INTELLIGENCE

Cloud. Automation. AI.
Connected with purpose.

Cloud, code and operations belong in the same conversation. I connect them so your team can build with confidence.

Let’s discuss your project
+HOW IT CONNECTS
Connected infrastructure, automation and intelligence An original isometric illustration of three floating glass system layers, connected by illuminated data paths.
AI SYSTEMS 03 AUTOMATION 02 CLOUD INFRA 01

One connected stack. Built to work together.

01Cloud & infrastructure

From Linux and container operations to repeatable deployments, migrations and infrastructure as code. Built with the operating team in mind.

AWSLinuxDockerLXD / LXCTerraformCI/CD
02AI & workflow automation

Practical AI assistants, department-scoped knowledge and business integrations. Clear access boundaries, observable workflows and human handoffs.

n8nPythonOpen WebUIRAGQdrantMCP
03Agent engineering & observability
WORKING KNOWLEDGE

I’m familiar with the following tools for AI agents, application workflows and LLM observability.

Agent tools

OpenClawHermes AgentHermes A2AHermes Kanban

Application workflows

LangChainLangGraph

Tracing & evaluation

Langfuse
04Reliability & recovery

Monitoring that makes issues visible. Backups with understood restore paths. Documentation and handover that make a system useful after launch.

ZabbixUptime KumaKopiaS3 / WasabiCloudflare
IN MY TOOLKIT
aws⌣DockerTerraformn8nPythonCloudflareOpen WebUI
03 /THE PERSON BEHIND THE SYSTEMS
Nguyễn Quốc Khang in a navy suit
BASED IN VIETNAM.
BUILDING BEYOND BORDERS.
↗

An engineer.
A thinking partner.

I’m Khang, a DevOps engineer based in Ho Chi Minh City. I work where infrastructure, automation and business operations meet.

I like making complex systems easier to run: fewer manual steps, clearer signals and a recovery plan you can actually use. The work is only complete when the people behind it can operate it confidently.

DevOps Engineer · ITMS VietnamDEC 2023 — PRESENT
01 / OBSERVE

Make it visible.

Turn logs, health checks and alerts into informed decisions.

02 / REPEAT

Make it repeatable.

Version the infrastructure. Automate the routine. Document the path.

03 / RECOVER

Make it recoverable.

Build for ordinary days. Plan for the difficult ones.

04 /LET’S BUILD SOMETHING THAT WORKSOpen to remote collaboration

Let’s solve your
next challenge.

Tell me about your project, what you need to improve and where you’d like to start.

Ho Chi Minh City, VietnamGMT+7 · Remote collaboration

START A CONVERSATION

Tell me what you’re building.

All fields are required.

Verification will load when you reach this form.

Your details are used to respond to this enquiry. Privacy policy