From EE Student to
AI Co-Founder
Born M. Shayan (2004). In 25 months I went from zero web development experience to co-founding Leaveit2AI and shipping three live products under one umbrella — bootstrapped, no ad spend.
In my third year of Electrical Engineering at Sukkur IBA University, I received an email about the Google Gemini AI Hackathon. I had taken courses in Java, Python, and Arduino — but I had never built a real product.
Working with Dr. Gul Sher Baloch, we built PoultaVet AI — an AI-powered poultry disease management system. Our team connected with Poulta Inc., which led to a Kedarobiz sales role and taught me real customer relations. By mid-2024, Dr. Gul Sher and Dr. Muhammad Yameen Sandhu invited me to co-found Leaveit2AI.
The learning curve was steep. I started with Excel sheets as databases, then rapidly progressed to MySQL, Redis, proper architecture, Linux deployment, DNS management, and Nginx configuration — all learned through production necessity, not curriculum.
By the time I built GradexAI and later GulfMedExams, I was no longer learning by breaking things. I was engineering systems from first principles — discoverability stacks, behavioural product loops, custom analytics, and lean serverless architecture. I graduated with a Silver Medal (CGPA 3.63) and the highest FYP marks in my batch.
The biggest lesson: English became my favourite programming language. As AI tools matured, I transitioned from assistance to full control — using AI as a pair programmer, not a crutch. Custom solutions, built from scratch. No boxed frameworks that limit what you can see.
From Excel DBs to Production Systems
How I went from zero web experience to architecting discoverability stacks in 25 months.
First Real Product — PoultaVet AI
- Built for the Google Gemini AI Hackathon with Dr. Gul Sher Baloch
- First exposure to AI APIs and product thinking outside a classroom
- Excel sheets as the first database — learned by breaking production
- Connected with Poulta Inc.; Kedarobiz sales role taught customer foundations
Co-founding Leaveit2AI
- Built the initial AI tools directory with a tag-based recommendation engine
- Team shipped AI PPT generator, AI Flashcard app, and Designs (AI clothing tool)
- Rapid upgrade: proper databases, Flask architecture, Redis, real deployment
- Business instincts from Kedarobiz shaped how we thought about product-market fit
GradexAI — FYP & Flagship
- Built GradexAI as Final Year Project — proper architecture from day one
- MySQL + Redis + Celery + Flask + Next.js (migrated from Jinja2)
- Secured highest FYP marks in the entire batch at Sukkur IBA University
- Dr. Ghulam Mujtaba (Regis University, USA) joined as advisor
- Now serving 5k+ users across 10+ institutions — 90% grading time saved
Graduation · Silver Medal · National Recognition
- Graduated Spring 2025 — Silver Medal, CGPA 3.63, Highest FYP marks in batch
- Ignite AI Wrapper — Regional Winner
- Selected for national-level showcase
- Leaveit2AI scales to 1M+ search impressions, DA 39, 10k+ professionals, 32% USA traffic
- Built discoverability infrastructure: SEO/AEO/GEO, GSC, Bing Webmaster, custom analytics
GulfMedExams — First Product Post-Graduation
- First new Leaveit2AI product after Silver Medal and Ignite recognition
- Built Gulf medical licensing exam prep: SMLE, DHA, DOH, MOH, QCHP and more
- Applied behavioural psychology to study sequencing — not just question dumps
- Registered users from 26+ countries, 60,000 MCQs solved within 30 days of launch
- Full AI Ops admin system with advanced analytics; zero paywall on core features
Co-Founder · Leaveit2AI Product Suite
- Leading product engineering across Leaveit2AI — GradexAI, GulfMedExams, AI tools platform
- Specialising in serverless launches, NFC systems, AI integrations, custom analytics
- Full-cycle builder: architecture → SEO/AEO/GEO → behavioural loops → launch → scale
- Building custom systems from scratch — session + JWT decoupling, stack migrations across all products
One Company. Three Live Products.
Leaveit2AI is the umbrella. GradexAI, GulfMedExams and the AI tools platform are its product lineup — bootstrapped, no VC funding, no ad spend.
Leaveit2AI
AI tool discovery platform. 1,000+ tools indexed, RAG-based assistant, comparison engine, taste profile system, and behaviour-informed discovery loops. Serving 10,000+ professionals across 32% USA traffic — organic, zero ad spend.
GulfMedExams
Gulf medical licensing exam prep — SMLE, DHA, DOH, MOH, QCHP and 4 more. Behavioural psychology baked into study sequencing. Full AI Ops admin. Free, no paywall.
- 26+ countries · registered users in month one
- 60,000 MCQs solved within 30 days of launch
- 8 major Gulf exams · 6 specialties · 10,000+ questions
- Competing against VC-funded platforms, bootstrapped
GradexAI
AI educational assessment — grades 30 assignments in the time it takes to do 3. Personalised feedback, 90% LLM consistency. Highest FYP marks in batch.
- 5k+ active users across 10+ institutions
- 15,000+ AI-powered assessments processed
- 90% grading time saved · 86% teacher alignment
- Flask + Celery + Redis + Next.js — migrated from Jinja2
The People Behind Leaveit2AI
Founded by two faculty researchers at Sukkur IBA University with combined decades of international academic and industry experience.
- Assoc. Professor, Electrical & Electronic Engineering · Sukkur IBA University
- Associate Editor · ETASR — Engineering, Technology & Applied Science Research, Greece
- 20+ publications · 349+ citations · CNN, Signal & Image Processing, NLP
Convolutional Neural Networks · Image Processing · Applied ML · NLP
- Assoc. Professor, Sukkur IBA University · 12+ years teaching, research & administration
- PhD & Postdoc · University of Leeds, UK (2011–2016) · Postdoc · Gdansk University of Technology, Poland (2021–22)
- IEEE Member · HEC PhD-approved Supervisor · Multiple HEC & PSF funded research projects
Microwave Material Processing · Passive Component Design · RF Sensing & Spectroscopy
Design Principles That Actually Ship
Custom Over Boxed Frameworks
Frameworks like Django impose opinions that limit what you can see and control. I build from scratch — not because it's harder, but because lower-level control produces better products. You can't build taste profile logic or custom discovery loops inside someone else's abstraction.
Discoverability as a Compounding Asset
SEO, AEO, and GEO are not marketing tactics — they're product infrastructure. Every content signal, schema tag, and GSC insight compounds over time. Leaveit2AI went from 200 to 600+ clicks/month in two weeks after 16 months of flat — organic, no spend.
Behaviour First, Then Architecture
Applied behavioural psychology shapes how I design every flow. Taste profiles, recall-first study sequencing, engagement loops — these aren't UX tricks, they're product fundamentals. Custom analytics tell me why users stay or leave, not just that they did.
Lean from Day One, Scalable by Design
Serverless and zero-to-low-cost launches aren't shortcuts — they're the right architecture for early-stage products. When you build lean, you're forced to make intentional decisions. Every line of infrastructure I wrote for GulfMedExams was deliberate.
The Journey Continues
If you're a healthcare professional preparing for Gulf exams, a founder thinking about AI products, or a teacher drowning in grading — I've built something for you. Use it. Tell me what's broken.