Community

Study groups, questions answered, mentors and events. Almost nobody finishes a course alone — this is the part that makes it possible.

Study groups

Learn on a schedule, with other people

A group works through one subject together over six to ten weeks with a volunteer facilitator. Free, and you do not need to attend every session. Groups start as each subject opens — register interest now and we will place you.

Subject 01

Python

Eight weeks, twice a week, one hour a session. Opens with the subject — register your interest now.

FoundationsMax 25 people
Subject 02

Mathematics for AI

Six weeks, Saturdays. Worked by hand on paper before any code, which is the only way it sticks.

FoundationsMax 30 people
Subject 03

Machine Learning

Ten weeks with graded exercises and a reviewed project at the end. Twice a week, evenings.

CoreMax 20 people
Subject 06

Natural Language Processing

Six weeks, ending with an open dataset that the group collects and publishes together.

AppliedMax 20 people
Subject 08

Agentic AI

Eight weeks, project-based. You ship an agent that does a real job and present it to the group.

AppliedMax 15 people
Start one

Facilitate a group yourself

You do not need to be an expert — a facilitator keeps the schedule and asks the questions. We provide the plan and the slides.

Questions and answers

Asked often enough to answer here

Do I need a laptop, or is a phone enough?

A mid-range Android phone gets you through the foundations, including most of Python, using Google Colab in a browser. From subject 03 onwards a laptop makes life much easier — mainly because of the keyboard and the screen, not the processing power, since the heavy work happens on Colab's servers. If a laptop is the only thing stopping you, write to us; we keep a small equipment list.

Do I need to know English?

Some, yes — the courses are taught in English, and so is every piece of documentation and research in this field. But the English you need is technical rather than conversational, which is a far smaller target than it sounds: a few hundred recurring words, plus the confidence to read a manual slowly. We keep the language plain, define every term once, and never use an idiom where a plain verb will do.

How much mathematics do I really need?

Less than people fear, and more than the marketing suggests. Subject 02 exists precisely for this: it covers the linear algebra, calculus and probability that actually appear inside models, and it names what you can skip. For machine learning you need comfort with algebra and an idea of what a derivative means. For deep learning you need to be genuinely at ease with matrix multiplication and the chain rule.

How long until I can get paid work?

We will not promise a number, because it depends on your starting point and how many hours a week you have. What we can say: the people who reach paid work are the ones who finish projects and publish them, not the ones who watch the most lessons. Two finished projects on GitHub with a clear write-up is worth more to an employer than ten certificates.

Is there a certificate?

We will issue a completion record for each subject, and we are honest about its weight: it is not accredited and no employer knows our name yet. Your project portfolio is the credential that works. We would rather help you build that than sell you a PDF.

Can women take part safely?

Yes, and it is a priority. All courses can be followed entirely at home with no attendance requirement. Some study groups are women-only with women facilitators. In any group you may use a first name only, and no photograph or camera is ever required. If a group does not feel safe, tell a facilitator or write to us directly and we will act.

I found an error in a lesson.

Thank you — please tell us. Open an issue on GitHub if you are comfortable with that, or send an email. Corrections are credited in the lesson notes, and finding a real error in the curriculum is one of the best first contributions you can make.

Learners

Profiles open with the first subject

A public profile will be optional and useful: it helps you find study partners, and it gives an employer something to look at. First name only is fine, and no photograph is ever required.

What a profile will show. The name you choose, roughly where you are, which subjects you have finished, and links to any projects you have published. Nothing else — no age, no contact details, no photograph unless you add one yourself.

What it will never show. We will not publish a full name, a photograph or a location more precise than a city without your written consent, and you can delete a profile at any time by sending one email. For learners for whom being visible online carries real risk, a profile is not required to study, to join a group, or to be matched with a mentor.

Register your interest

Events and webinars

What is on, every week

Recurring sessions, so you can plan around them. All times are Kabul time (UTC+4:30) unless stated, and everything is recorded. Sessions begin as the first subjects open.

Recurring AAII events
WhenSessionFor whom
Saturdays 19:00Live lesson and open questions — one topic, then anything you want to askEveryone
Sundays 18:00 UKMachine learning study group — exercise reviewSubject 03 learners
Tuesdays 20:00Code clinic — bring a broken program, leave with it workingEveryone
First Thursday monthlyGuest talk — an engineer or researcher on what they actually doEveryone
Last Friday monthlyProject showcase — five learners present, ten minutes eachEveryone

Mentorship

One hour a month changes someone's year

We pair learners with working engineers and researchers, many of them Afghans abroad. A pair meets for an hour a month for six months. The mentor does not teach the course — they answer the questions a course cannot: which path to take, whether a project is good enough, how to read a job advertisement, what to do after a rejection.

Demand exceeds supply, so priority goes to learners who have finished at least one course and published a project.