Online · Applied AI engineering · Urdu and English
Learn AI by building things that run.
Not tutorials. Programs where every module ends in a project that ships to your GitHub — from zero computer science, or in three months if you already code.
You finish, you have watched forty hours, and you still cannot start a project from an empty folder. We built this the other way around.
01
Projects, not playlists
Every module ends in something deployed and public. Nine of them by the end of the diploma. The portfolio is the qualification — the certificate is a formality.
02
Real data, mess included
No Titanic. No Iris. Datasets with missing values, leakage traps and lies in them, because that is what you will be handed on day one of a job.
03
We publish what we do not teach
Backpropagation derivations, foundation models from scratch, Kubernetes. A course that claims everything teaches nothing.
04
Entry is assessed, honestly
If you are not ready for the bootcamp we say so and point you at the diploma. Carrying an unprepared student hurts them and slows the cohort.
What you build
Six things that go on your GitHub, not on a certificate.
Each is a module deliverable. Each is defensible in an interview, because you will have to defend it in front of the cohort first.
Computer Vision01
A vision system on live video
Collect and label your own data, train a detector, run it on an RTSP stream. Measured mAP and FPS.
Agentic AI02
An agentic application
RAG over a real corpus, tool calling, guardrails, and a measured cost per query.
NLP03
A fine-tuned language model
LoRA on your own data, including Urdu and bilingual text, with a baseline and error analysis.
MLOps04
A deployed, monitored service
Dockerised, CI/CD, model registry, drift alerts that actually fire.
Production ML05
A model that survives review
Leakage-audited, cost-based threshold, SHAP explanations, limitations stated in writing.
Data06
A pipeline that runs unattended
Ingest, clean, store, serve and chart — on a schedule, surviving failure.
Sprints
Need one thing, not eight months?
Seven short courses, three to five weeks each, one project each. Together they are the three-month bootcamp.
Yes. Every program is taught online, so you can join from anywhere in Pakistan or abroad. You need a laptop and a stable internet connection; the GPU work runs in the cloud, so you do not need an expensive machine.
Do I need a computer science degree?
No. The eight-month diploma starts at the terminal and assumes nothing. The three-month bootcamp does assume you already program — entry is assessed, and if you are not there yet we will tell you and point you at the diploma.
What language are classes taught in?
Urdu and English, mixed the way engineers in Pakistan actually work. Explanation and discussion happen in whichever makes the idea land fastest. All material, code, documentation and assignments are in English, because that is the language the job is in — and every technical term is taught in English from day one.
How much time does this take outside class?
Six to eight hours a week for the diploma, eight to ten for the bootcamp. This is not optional. A student who does only the contact hours will finish the course and struggle in interviews — we say this before you enrol, not after.
Do I need a GPU?
Not to start. Colab's free tier covers most of the teaching. The computer vision and fine-tuning modules need more, and we tell you exactly what to budget before those modules begin.
Do you guarantee a job?
No, and be careful with anyone who does. What we guarantee is a portfolio of deployed, documented projects you can defend in an interview, plus interview practice in the final module.
Is there a certificate?
Yes, but the portfolio is the qualification. Nobody has ever been hired on the strength of a certificate from a company they had not heard of.
Next cohort starts soon. Places are limited to 25.
Applications take ten minutes. We reply within three working days, either way — or just message us on WhatsApp.