ABOUT ABDUL QADEER

I build systems that work in production — not demos.

I'm Muhammad Abdul Qadeer — an AI/ML engineer, software & web application developer, and automation engineer. I founded Qadeer Automations, an AI automation agency serving clients in the UAE, UK, USA, Australia and Pakistan.

Mathematician by training. Engineer by building.

I'm based in Jatoi, Muzaffargarh, in Punjab, Pakistan, and I work remotely with founders and businesses worldwide. My route into technology wasn't a straight line: I studied mathematics — a BS at Ghazi University and a BSc at Bahauddin Zakariya University, Multan — then spent about five years in HR coordination and virtual assistance before teaching myself AI and machine learning from 2024, learning from online teachers and, mostly, by building systems that had to work for real users.

That background shapes how I work. Operations first, models second: a prediction only matters if it changes a decision somebody actually makes, and an automation only matters if it survives contact with a real business process. Everything on this site — the services, the systems, the numbers — is stated the way I would want it stated to me: plainly, with the limitations included.

What I build

Selected systems I've built

▸FloodGuard AI

A flood-susceptibility tool for South Punjab planning: a Random Forest trained on a 56,615-settlement index, 79 historical flood events (2006–2025) and rainfall, elevation and river-distance features. A public planning interface — not a live forecast or official warning. View live ↗

▸Pulse Demand API

Product-level retail demand forecasts served through a FastAPI backend with a dashboard. A demonstration system — a client build is retrained and validated on that business’s own data. View demo ↗

▸Smart School Pro

A complete school management system — students, attendance, staff roles and records — built with React, TypeScript and Supabase, and deployed live. View live ↗

▸SafeRoad AI

A YOLOv8 detection pipeline that turns road footage into reviewable, locatable findings for maintenance planning. A detection system for review workflows — not an autonomous road-safety decision maker.

How I work

1

Validate before promising

No accuracy figure is quoted before a baseline exists and the model has been tested on data it never saw in training. If the signal isn't there, I'll say so — before you've paid for a build.

2

Document the limitations

Every delivery includes what the system can't do, where it's blind, and what would improve it. A model with documented limits beats a confident black box.

3

Working software, not demos

The deliverable is a deployed system your team can use — an API, a dashboard, an automation — with documentation and a handover you can act on.

4

Smallest useful phase first

Start with the smallest phase that proves value, judge it on real data, then scale. You commit to scope in writing, one phase at a time.

Background & credentials

🎓Education

BS Mathematics — Ghazi University. BSc — Bahauddin Zakariya University (BZU), Multan. The mathematics foundation underpins the modelling work: statistics, probability and optimisation, applied daily.

🏅Recognition

Micro1 AI/ML "Outstanding Performance" certificate, awarded for machine learning and data science work assessed on the Micro1 platform.

🧭Before AI/ML

Around five years in HR coordination and virtual assistance — screening, scheduling, records and operations for remote teams. It's why my systems are designed around how businesses actually run.

🛠Core stack

Python · XGBoost · scikit-learn · FastAPI · Streamlit · React · TypeScript · Supabase · PostgreSQL · n8n / Make · YOLOv8 · SHAP · Groq / Gemini APIs.

Let's talk about your project

Book a free 30-minute review or message me on WhatsApp. Tell me the decision you're trying to improve — I'll tell you honestly what it would take.

Book a Free Review WhatsApp Me