Machine Learning Fundamentals
What machine learning actually is, when it beats rules, and how to train and honestly evaluate your first model — the free entry to the Machine Learning course.
That was a one-minute preview. The full course video is for signed-in learners. Sign in to watch it here — or create a free BvLogic ID; it takes a minute, and this course is added to it.
What you'll be able to do
- Understand when learned patterns beat written rules, what ML genuinely cannot do, and the honest shape of the field.
- Prepare a dataset for learning — features, labels, encoding, and the cleaning that decides more than the algorithm does.
- Train, inspect and use two classic models with scikit-learn — and understand what 'training' actually does.
- Hold out test data, read accuracy against the base rate, use precision and recall, and recognise overfitting — the discipline everything else depends on.
- Map the real local applications and roles, and choose beginner projects that build honest skill rather than tutorial-copying.
- Understand what happens after training — saving, serving, monitoring, ethics — and see the road through the Machine Learning course.
Machine Learning
Machine Learning
This free course is the introduction. The programme is taught by a trainer, marked by the academy, and carries the syllabus in full — the part an employer asks about.
See the full syllabus and feeCourse content
- 1 Rules versus learning: what ML actually is Understand when learned patterns beat written rules, what ML genuinely cannot do, and the honest shape of the field.
- 2 Data and features: what the model actually eats Prepare a dataset for learning — features, labels, encoding, and the cleaning that decides more than the algorithm does. Sign in to open
- 3 Training your first models Train, inspect and use two classic models with scikit-learn — and understand what 'training' actually does. Sign in to open
- 4 Honest evaluation: the lesson that is the course Hold out test data, read accuracy against the base rate, use precision and recall, and recognise overfitting — the discipline everything else depends on. Sign in to open
- 5 ML in Pakistan: where it is actually used, and the projects that teach Map the real local applications and roles, and choose beginner projects that build honest skill rather than tutorial-copying. Sign in to open
- 6 From model to use — and what the paid course adds Understand what happens after training — saving, serving, monitoring, ethics — and see the road through the Machine Learning course. Sign in to open
Course lab
Graded lab The dropout model, trained, evaluated honestly and deployed Open the lab →Who this course is for
- Python learners who want to train and test a first model.
- Analysts who keep hearing 'machine learning' and want to know when it beats a written rule.
- Anyone considering the paid Machine Learning course.
What the 6 lessons cover
- Rules versus learning: what ML actually is. Machine learning finds the rule from labelled examples. Use it when the rule is real but too subtle to write down, and write the rule yourself when a clear one exists.
- Data and features: what the model actually eats. A table of features and one label, built on a student-dropout example. Choosing features is the craft, and features that leak the answer are the danger.
- Training your first models. The same four lines of scikit-learn train a decision tree you can read and a second model that teaches the opposite trade-off.
- Honest evaluation: the lesson that is the course. Never grade on training data: hold out a test set so the train-test gap reveals overfitting, beat the always-predict-the-majority baseline, and use precision and recall when classes are imbalanced.
- ML in Pakistan: where it is actually used, and the projects that teach. Credit scoring, fraud detection, telecom churn, e-commerce demand and agriculture pilots. Most people reach ML roles through analyst or Python developer jobs first.
- From model to use — and what the paid course adds. Saving a model, serving it behind a small API, and monitoring it as the world drifts, plus the ethics of models that make decisions about people.
Try the first exercise (in Urdu)
انہیں rule-بمقابلہ-ML میں بانٹیں اور ہر ایک کی وجہ ایک line میں لکھیں: (1) 30+ دن سے بقایا فیسوں کی نشاندہی؛ (2) پیش گوئی کہ کون سی enquiries داخلوں میں بدلیں گی؛ (3) 5 ناکامیوں کے بعد login بند کرنا؛ (4) scan کیے گئے forms سے ہاتھ سے لکھے roll numbers پڑھنا؛ (5) فیصلہ کہ کلاس میں اول کون ہے۔ دو rules ہیں، دو ML، ایک جال ہے (مثالیں موجود نہیں / pattern سیکھا نہیں جا سکتا) — جال ڈھونڈیں۔
Using this in Pakistan
- Indeed's range for junior data analysts in Pakistan is Rs 38,190 to 101,116 a month, averaging Rs 62,142, from only 12 reported salaries. Source
- Few Python adverts show pay: a Lahore Python backend internship offered Rs 15,000 to 20,000 a month, and none of the Faisalabad Python or data adverts seen on 30 September 2026 listed a salary. Source
Questions
Is it really free?
Yes. No payment and no card; lesson 1 needs no account.
Is there a certificate?
Yes, when you complete every lesson.
Can I learn in Urdu?
Every lesson has an Urdu exercise alongside the English reading.
How long does it take?
About 18 minutes of reading, plus the exercises.
What comes next?
The paid Machine Learning programme, in Faisalabad and live online.
Related free courses
- AI Agents · Urdu video
- AI Ethics, Disclosure and Safety · Urdu video
- AI for Everyday Work · Urdu video
- AI Fundamentals · Urdu video
- AI Image Generation · Urdu video
- AI Video Generation · Urdu video