Available for remote roles and project work, any time zone
Ahsan Tariq
Full Stack Developer and n8n AI Automation Expert
I turn repetitive work into systems that run on their own.
I trained as an industrial engineer, so I map your process on paper before I automate it. Then I build it in layers you can test, trace and hand over. And every system I build asks a person before it does anything risky.
- 7 h to 45 min Daily lead collection time after my screenshot scraper went live
- 10 systems 8 n8n automations and 2 SaaS products, every one shown below with real screenshots
- 10 years Building systems: factory lines, a funded hardware startup, now software
Why teams can rely on me
Five promises I keep on every project. Each one comes with proof you can check on this page.
Start with a short call-
I plan before I build
Every project starts with a written requirement document (PRD). Think of a house: drawings first, then bricks. You see exactly what will be built, and what it will cost to run, before a single workflow exists.
ProofInvoice Sync Match had its PRD and data model finished before the first n8n engine was built.
-
AI does the work. A person makes the call.
AI is great at reading, sorting and drafting. Money, customers and legal documents still need a human. So my systems prepare everything, then wait for a click.
ProofInvoice Sync Match posts 0 invoices to QuickBooks without someone pressing Approve. DenialDesk makes 0 automatic insurance submissions.
-
I measure the result in hours you get back
A system is only worth it if someone gets their day back. I track the before and the after.
ProofReal estate lead collection went from 7 hours a day to 45 minutes. That is 6 hours 15 minutes back every working day, or about 125 hours a month.
-
One person owns the whole stack
Website, dashboard, n8n backend, database, AI, server and SSL. You don't need three vendors pointing at each other when something breaks.
ProofMy own n8n runs self-hosted on a Linux server with Docker, PostgreSQL and Caddy. Both SaaS products below run on infrastructure I set up.
-
I have already built and run a real company
As co-founder and CTO of Motto Vest, I took a motorcycle airbag vest from a 2019 prototype to a product people buy. A safety product cannot fail silently, and I bring that same discipline to software.
Proof1,350+ units sold since October 2025, a $100,000 seed round, and 3 national awards. Details in my journey below.
What could automation give back to your team?
Move the sliders to match one repetitive task in your business, like copying details from emails or web pages into a sheet.
The estimate uses my real result: a 7 hour daily task that now takes 45 minutes, which is 89% less time. Every task is different, so treat this as a starting point for a conversation, not a promise.
- 18 full 8 hour working days
- 1,714 hours in a year
- 0.9 of a full-time person's month
SaaS product 1 of 2
Invoice Sync Match
Accounts payable automation for a group of 6 companies. It reads supplier invoices, checks them against what was ordered and what arrived, and lets a person approve before anything reaches QuickBooks.
The problem, in plain words
Picture 6 companies under one owner. Each gets supplier invoices by email every day. Someone opens each PDF, compares it with the purchase order and the delivery note, then types it into QuickBooks.
One wrong number means paying a supplier too much. One file filed under the wrong company means messy books for all of them.
What I built
The system reads the email, pulls every line out of the invoice with AI, finds the matching purchase order and delivery note, and checks that all three agree. This is called a 3-way match.
If something is off, it shows exactly which line and by how much, and drafts a correction email. A person still clicks Approve before anything is posted.
- 6separate companies, each with its own inbox, suppliers and QuickBooks account. Nothing crosses between them.
- 4n8n engines, each with one job: read purchase orders, read invoices, check the match, carry out the decision.
- 80%auto-match rate on the dashboard shown below. In simple terms, 4 of every 5 invoices only needed a quick look.
- 0postings without a human click. Every read, check and decision lands in an append-only audit trail.
Try it: run a 3-way match
Pick a scenario and press Run. This mirrors a real review screen from the product.
- Invoice email from Kestrel Fasteners Ltd received, PDF saved
- AI reads 3 invoice lines into a fixed format
- Purchase order PO-26-4182 found, same company only
- Delivery note confirms the quantities that arrived
- Every line compared across all 3 documents
| Item | Ordered | Arrived | Price on order | Price on invoice | Check |
|---|
Values come from the test company in the screenshots. Prices in GBP, VAT 20%.
How the pieces connect
Each engine does one job, so a problem in one place never takes the whole system down. The glowing lines show the path an invoice takes.
Real screenshots
SaaS product 2 of 2
DenialDesk
When an insurer refuses to pay a medical claim, DenialDesk turns the denial letter into a ready-to-review appeal packet. I built the backend automation.
The problem, in plain words
A US clinic sends a claim. The insurer says no and sends back a letter full of codes. A billing specialist has to decode it, work out what went wrong, gather evidence and write an appeal. Multiply that by hundreds of claims.
How the backend works
- 1A specialist uploads the denial letter on a web form.
- 2n8n pulls the text out of the PDF.
- 3GPT-4o explains the denial reason and maps the appeal steps.
- 4The answers fill a Google Docs template through the API.
- 5The document becomes a PDF packet, ready for a person to review.
- 5 to 10 mintarget time for a full case workup
- 0automatic submissions to insurers. A qualified person always reviews first.
The packet template the system fills
Real screenshots
n8n automation systems
Automations running real work
Each card is a system I designed and built. Every image is the real n8n workflow, not a mockup. Open any card to see the problem, how it works, and what changed.
Under the hood
n8n is the backbone, but the reliable parts are often a few lines of careful code. These are simplified examples of patterns from my builds. Client code stays private.
- Clean data in, clean data out. Duplicates and bad emails are caught before they reach your sheet, and anything unclear is flagged for a person.
- Companies never mix. Every query is locked to one company, and every decision is written to a log that can't be edited.
- AI answers in a fixed shape. The model must return exactly these fields, so a missing total is caught instead of guessed.
- Hosted on servers I control. n8n, PostgreSQL and automatic SSL, restarted on their own if anything stops.
No vibe coding. The same 7 steps, every project.
I trained as an industrial engineer, so I map the process before I automate it. Every system starts on paper and is built in layers you can test, trace and hand over.
Why it matters to you: fewer surprises, a clear price, and a system your own team can understand after I hand it over.
-
1
Write the requirement document
What happens today, what should happen, and who approves what. Written down before anything else.
You get a clear document to read and sign off.
-
2
Find the core problem
I look for the real bottleneck, not just the symptom. The slow part is often not where people expect.
You get agreement on what success looks like, in numbers.
-
3
Choose the tech stack
The right tools for your budget and team. Sometimes n8n alone, sometimes custom code and a database, often both.
You get a stack you can afford to run and maintain.
-
4
Design the architecture
Every engine, data table and decision point is mapped, including what happens when an API fails or a file is missing.
You get a diagram of the whole system.
-
5
Lock the architecture
We review it together and freeze it, so the build doesn't drift halfway through.
You get no surprise rebuilds.
-
6
Build in tested layers
Intake, checks, AI and actions are built and tested one layer at a time.
You get progress you can test at every layer.
-
7
Log everything in GitHub
Every change is committed and documented, so your team can follow the work and anyone can take over later.
You get a full history and a clean handover.
Ten years of building systems
First production lines, then a company, now software. The tools changed. The job stayed the same: make a process run better.
Motto Vest, co-founder and CTO
An AI-powered airbag vest for motorcycle riders. It senses a crash and inflates in milliseconds. I led the technology: sensors, firmware, the deployment logic, mechanical design, manufacturing and supply chain.
Today the team I built runs it day to day, and I set direction and policy. That is why my full working time now goes into software.
- 1,350+units sold since October 2025
- $100Kseed round from Sanabil Orbit Ventures, Saudi Arabia
- 3national awards, including the Presidential Award
- 2countries: made in Lahore, registered in ADGM, UAE
- 2012
BSc Industrial and Manufacturing Engineering
UET Lahore, 2012 to 2016. Where I learned to map a process before changing it.
- 2016
Industrial Engineer, RK Gears
Gear production lines, quality control, process improvement and design software supervision.
- 2018
Product Development Specialist, Dimen Draw
Machinery design, plant layout and engineering proposals, alongside freelance design work in SolidWorks and ANSYS.
- 2019
Motto Vest begins
First raw prototypes of a motorcycle airbag vest.
- 2021
MS studies, Mechanical Design Engineering
UET Lahore.
- 2022
Motto Vest incorporated, CTO. Shell Tameer Award.
A $3,000 grant to develop the product.
- 2023
Pakistan National Innovation Award
An $8,000 grant, awarded by the Prime Minister of Pakistan.
- 2024
Presidential Award, National Idea Bank. The move into software.
Started building AI automations and SaaS systems with n8n, JavaScript and PostgreSQL.
- 2025
$100,000 seed round
From Sanabil Orbit Ventures, Saudi Arabia. Sales began in October.
- 2026
GCC expansion and Lead Sync Intelligence
MV Safety Solutions Limited registered in ADGM, UAE. I launched Lead Sync Intelligence, my AI automation practice, where I lead the technical work.
Tools I ship with
Software is my focus now. The engineering background is why I think in processes, tolerances and failure points.
Frontend
- JavaScript
- HTML
- CSS
- Custom dashboards
- Websites
Automation and backend
- n8n (self-hosted)
- Node.js
- Webhooks
- REST APIs
Data
- PostgreSQL
- Cloudflare R2
- Google Sheets
- Airtable
AI
- OpenAI APIs
- Structured outputs
- Vapi voice agents
- Claude Code
Infrastructure
- Docker
- Linux VPS (Hetzner)
- Caddy
- DNS and SSL
- Git and GitHub
Engineering background
- Production management
- Ergonomics and work study
- Computer integrated manufacturing
- CAD/CAM
- SolidWorks
- ANSYS Workbench
- CFD
Certifications
- Advanced n8n Automation
- Pressure Vessel Design
Have a task that eats your team's week?
Describe it in two or three lines. I will reply with how I would automate it, which tools it needs, and what a first version would look like.
