Mads Alhamid

Portfolio · 2026 — Present

Systems that turn hours into seconds.

A selection of end-to-end automation and AI projects, designed and built independently for recruitment, customer service, marketing and sales.

AI-Powered CV Screening & Recruitment Automation

80%

less manual screening effort*

13+ hrs

saved per day at 200 CVs*

1 min

per CV instead of 5 minutes*

01

HR Automation · AI Screening

AI-Powered CV Screening & Recruitment Automation

An intelligent screening workflow built on self-hosted n8n with Google Gemini, Google Forms, Drive, Sheets and Gmail. It collects applications, reads every CV, scores each candidate against the role and replies automatically.

  • Collects applications through Google Forms with custom logic and automatic CV duplicate detection.
  • Downloads each CV from Google Drive and extracts its content from PDF automatically.
  • Scores every candidate with a Gemini AI agent against the job specification, returning structured results.
  • Records each result in Google Sheets and routes candidates by qualification criteria.
  • Sends personalised acceptance or rejection updates through Gmail.
n8n (self-hosted)Google GeminiGoogle FormsGoogle DriveGoogle SheetsGmail

Proof of work

Live · Published on n8n

I come from an HR and legal background, so I know how much time is lost to reviewing CVs, tracking candidates and following up. This is the system I once wished I had.

Published n8n workflow for AI CV screening: form submission, CV extraction, Gemini scoring and Gmail feedback
The live n8n workflow, published and running.
  1. Form submission
  2. CV extraction
  3. AI scoring with Gemini
  4. Sheets record + Gmail feedback

The problem

  • Manually reviewing hundreds of CVs
  • Applicant files scattered across drives and inboxes
  • Unstandardised, subjective evaluations
  • Manual database updates and repetitive follow-ups

Potential impact

Scenario: 200 CVs per day

Manual screening (5 min per CV)
16.7 hrs/day
AI-assisted screening (1 min per CV)
3.3 hrs/day
Reduction in manual effort
80%
Time saved
13+ hrs/day

*Illustrative estimates. The goal is to empower recruiters, not replace them.

Built for privacy and clean data

Runs on a self-hosted n8n instance for minimal data exposure, and duplicate applications are caught at intake so the candidate database stays clean.

I understand the problem because I've lived it. Now I engineer the solution.

AI Customer Service & Booking Automation

6.7 hrs

of work saved per day*

1,760 hrs

potential capacity per year*

≈US$14k

yearly labour capacity*

02

Conversational AI · WhatsApp Booking

AI Customer Service & Booking Automation

A WhatsApp AI agent for an auto workshop that answers customers from company SOPs, books service slots, records every booking and hands the chat to a human admin when needed. It is live and serving real conversations.

  • Answers customer questions based on the company's SOPs and rules.
  • Manages the booking schedule and checks availability to prevent double bookings.
  • Records and organises customer and booking data automatically.
  • Hands the conversation from AI to a human admin when needed, then reactivates the AI once the admin session ends.
  • Connects multiple business systems through automated workflows.
n8n (self-hosted)Google GeminiDockerREST APIGoogle SheetsWhatsApp (Fonnte)

Proof of work

Live · Published on n8n

What if a business could serve more customers without hiring more staff? This assistant takes over the repetitive side of customer service so the team can focus on the work itself.

Published n8n workflow 'Bengkel AI Agent & Human Handoff' with webhook, status checks, human handoff and the Gemini-powered AI agent
“Bengkel AI Agent & Human Handoff”: the live n8n workflow behind the chatbot.
  1. WhatsApp webhook
  2. Check bot / human mode
  3. AI agent with Gemini + memory
  4. Booking saved to Sheets
  5. Reply via WhatsApp

Screen recording of a real WhatsApp chat with “Bram”, the workshop's AI agent: greeting, service booking, admin handoff and the booking saved to the database.

Potential impact

Scenario: 100 customer conversations per day

Time saved per conversation
4 min
Working hours saved
6.7 hrs/day
Per month
147 hrs
Per year
1,760 hrs
Labour capacity at US$8/hr
≈US$14,080/yr

*Illustrative simulation, not measured results from a client deployment. Actual results depend on the level of automation, need for human intervention, operating costs and implementation.

Built for flexibility and control

Runs on self-hosted infrastructure, giving more control over deployment and supporting privacy-oriented data handling.

AI shouldn't only produce answers. It should run processes, connect systems and create business value that scales.

AI-Driven Marketing & Content Production Suite

≈80%

lower production cost*

0

talent, crew, studio or agency fees

Weekly

auto-generated campaigns possible

03

Generative AI · Creative Production

AI-Driven Marketing & Content Production Suite

A full creative production pipeline powered by generative AI, from brand logos and product photography to showcase videos and human-avatar marketing videos.

  • Professional logo concepts and brand assets produced in hours, not weeks.
  • Studio-grade product photography without a physical photoshoot.
  • Product showcase videos and human-avatar video marketing.
  • Content formatted for TikTok, Instagram Reels and YouTube Shorts.
HiggsfieldImage GenerationVideo GenerationAI AvatarsPrompt Design

Proof of work

Real campaign · HUMBLY beauty

For HUMBLY, a beauty brand, I produced the full marketing kit with AI: logo, studio-grade product photos and a human-avatar video spot. The avatar speaks naturally and sells the way a real presenter would, with no talent, camera crew, editor, location or agency involved.

HUMBLY beauty campaign: AI-generated perfume, cream and lipstick product photos and lifestyle visuals
HUMBLY campaign assets: product photography and lifestyle visuals, all generated with AI.
  1. Brand & logo
  2. Product photos
  3. Avatar video spot
  4. TikTok · Reels · Shorts
HUMBLY logo and monogram concepts generated on an AI image platform
AI logo & monogram exploration
HUMBLY AI-avatar lipstick promotion published on TikTok
Avatar spot live on TikTok

HUMBLY lipstick spot: a fully AI-generated human avatar with natural, conversational delivery, made with Higgsfield.

The problem

  • Hiring talent or influencers for every campaign
  • Camera crew, editing and studio rental costs
  • Agency retainers and weeks of turnaround
  • Slow, manual content output for new products

Potential impact

Traditional shoot vs. AI production

Talent / influencer fee
Not needed
Camera crew & editing
Not needed
Studio or location rental
Not needed
Advertising agency
Not needed
Overall production cost
≈80% lower

*Illustrative estimate based on my own projects using Higgsfield. Actual savings depend on scope and volume.

Ready to run on autopilot

The same pipeline can be automated to publish a new video spot every week, or trigger automatically whenever a new product is added.

Studio-quality campaigns, designed entirely with AI, at a fraction of the cost.

Smart CRM Lead Scoring & Sales Pipeline Automation

Top 20%

leads surfaced first

Weekly

automated follow-up reminders

100%

pipeline visibility

04

Sales Ops · Data Automation

Smart CRM Lead Scoring & Sales Pipeline Automation

An automated analysis system that ranks prospects by potential and delivers structured, prioritised leads to the sales team with built-in follow-up cadence.

  • Database analysis to identify and rank the highest-potential prospects.
  • Structured lead handoff delivered directly to the sales team.
  • Weekly timelines with automatic reminders for each follow-up.
  • A clear pipeline view so no warm lead goes cold.
n8n (self-hosted)WebhookGoogle SheetsGmailSlackGoogle Calendar

Proof of work

Built & tested on n8n

Sales teams lose deals when hot leads sit in a spreadsheet next to cold ones. This pipeline scores every new lead the moment it arrives and routes it to the right next step automatically.

n8n workflow 'CRM Lead Scoring & Sales Pipeline': webhook, read leads, prepare data, calculate score, check score, hot lead email and Slack alert, nurture list and calendar follow-up
“CRM Lead Scoring & Sales Pipeline” on my self-hosted n8n: executed successfully in 2.3 seconds.
  1. Webhook
  2. Read leads
  3. Calculate score
  4. Hot: email + Slack
  5. Warm: nurture + calendar follow-up

The problem

  • Hot and cold leads mixed in one list
  • Sales team chasing the wrong prospects first
  • Follow-ups forgotten or done too late
  • Manual copying between sheets, email and chat

Potential impact

What happens to each new lead

Time from new lead to score
≈2 seconds
Hot leads alerted to sales
Instantly
Warm leads followed up
Auto-scheduled
Manual data entry
None

*Run time taken from a test execution. Business results depend on lead volume and scoring rules.

Two clear paths for every lead

High-scoring leads trigger an instant email and a Slack alert to the sales team. The rest go into a nurture list with a follow-up automatically scheduled in Google Calendar.

Every lead gets a score, a next step and a reminder, without anyone touching a spreadsheet.

Have a process that eats your team's time?

Tell me how it works today. I'll show you how it could run on its own.

Start a project