Whatever you study, from systems engineering at a public university to economics, law or medicine at a private one, artificial intelligence is now part of your field, and your syllabus may not have caught up. This page is for university students in Peru who want a serious AI course alongside their degree: what to learn, what Peru's 2025 AI regulation asks universities to add, a lesson that shows why an algorithm can give two different answers on the same data, and how English fits in.
Teaching live since 2020 · 10,000+ students worldwide · all classes live and in English
What this is
Modern Age Coders offers an AI course for university students in Peru: live online classes taught in English, beside any degree. Students learn Python, the maths behind machine learning, supervised and unsupervised models, generative AI and agents, and they test every project on data the model has not seen. Mentors based in India teach a group of fifteen to twenty students or one student alone, at times that fit lectures and exams in Peru. It suits engineering and computer science students who want depth and students of other subjects who want AI they can actually explain.
Where to start
Three AI courses for undergraduates
The first is the broad foundation; the second goes into classical AI theory; the third is for students who already code and want generative AI and agents.
Why take an AI course while studying at a Peruvian university?
Because AI now touches almost every degree, and most programmes are still adding it. A focused course fills the gap without changing what you study.
Peru has a large university population. SUNEDU, the national university regulator, reported on 15 September 2021 that the country had 94 licensed universities, 46 public and 48 private, with 1.1 million university students across 5,500 programmes. The United States government's trade guide to Peru gave a figure of 1.2 million university students in its 2022 edition. The two counts come from different sources and years, so we print each with its own date rather than combine them.
Engineering and computer science
You may already code. What a course adds is depth in the machine learning your syllabus covers lightly: unsupervised learning, evaluation, generative AI and agents, with time to build projects properly.
Economics, business and science
You work with data already. A course teaches you to fit, test and question models yourself, in Python, instead of reading results someone else produced.
Law, health, education and the arts
You will meet AI in your profession. Understanding how a model learns and fails lets you judge its output, which is a different skill from using a chatbot.
Online access is rarely the problem. INEI's household survey for the second quarter of 2026 found that 98.0% of people in Peru with university education used the internet, and 97.3% of those aged 19 to 24. A laptop is the real requirement: only 36.4% of Peruvian households owned at least one computer in the same quarter, so many students rely on a university lab or a shared machine. Our classes work on any modest laptop; nothing in the first months needs a powerful graphics card.
What 2025 changed
What does Peru's AI regulation ask universities to teach?
It asks for AI content to be added progressively to related programmes. DS 115-2025-PCM, the regulation of Law 31814 published in El Peruano on 9 September 2025, names the disciplines.
Programmes named in article 13.3 of DS 115-2025-PCM
Programme area (as named)
In English
Our nearest course (our choice)
Ingeniería de sistemas
Systems engineering
AI and ML Masterclass; Search, Logic and Agents
Ciberseguridad
Cybersecurity
Generative AI: LLMs, RAG and Agents
Ingeniería electrónica
Electronic engineering
AI and ML Masterclass
Ingeniería de software
Software engineering
Data Structures and Algorithms; Codex and Claude Code
Ciencias de la computación
Computer science
Search, Logic and Agents; DSA
Tecnologías de la información
Information technology
Data Analysis
Matemáticas
Mathematics
Maths for data science; AI and ML Masterclass
Source: DS 115-2025-PCM, Reglamento de la Ley 31814, article 13.3, El Peruano, 9 September 2025 (the list also adds "demás disciplinas afines", other related disciplines). The right-hand column is our own mapping to our courses; it is not part of the regulation. Read on 9 November 2026.
What the text says
Article 13.3 says the government's digital transformation office "fomenta la incorporación progresiva" of AI content in those programmes, to strengthen graduates' profiles and respond to digital transformation. "Progressive" is the important word: the change happens over years, programme by programme.
Article 19 adds that the office may open public cloud computing resources for AI work to public bodies, researchers, teachers and undergraduate and postgraduate students. The same regulation gives public universities up to two years to implement several of its obligations.
What it means for you
If you are in one of those programmes now, AI content may arrive in your syllabus before you graduate, or after. A course alongside your degree means you do not have to wait for it.
We quote the regulation as published context, not legal advice. It does not endorse any course, ours included, and we are not connected with any Peruvian university, SUNEDU or the government.
A real lesson
Why can the same algorithm give two answers? k-means by hand
This lesson comes early in our unsupervised learning unit. Students write k-means from scratch in plain Python, run it twice on the same twelve points, and get two different results, one of them twenty times worse.
The scikit-learn user guide describes k-means as an algorithm that groups data by "minimizing a criterion known as the inertia or within-cluster sum-of-squares": the total squared distance from each point to the centre of its group. It repeats two steps, assign each point to its nearest centre, then move each centre to the mean of its points, until nothing moves. Our version below does exactly that in about twenty lines, with no library.
k-means from scratch, three starting positionsPython
# Twelve invented points: three tight groups on a plane.
points = [(1, 1), (1, 2), (2, 1), (2, 2),
(8, 1), (8, 2), (9, 1), (9, 2),
(5, 8), (5, 9), (6, 8), (6, 9)]
def dist2(a, b):
return (a[0] - b[0]) ** 2 + (a[1] - b[1]) ** 2
def kmeans(centroids, rounds=10):
for _ in range(rounds):
groups = [[] for _ in centroids]
for p in points:
nearest = min(range(len(centroids)), key=lambda i: dist2(p, centroids[i]))
groups[nearest].append(p)
new = []
for g, c in zip(groups, centroids):
if g:
new.append((sum(x for x, y in g) / len(g), sum(y for x, y in g) / len(g)))
else:
new.append(c)
if new == centroids:
break
centroids = new
inertia = sum(min(dist2(p, c) for c in centroids) for p in points)
return centroids, inertia
for name, start in [("start A", [(1, 1), (8, 1), (5, 8)]),
("start B", [(1, 1), (2, 2), (7, 5)])]:
cents, inertia = kmeans(start)
print(name, "->", [(round(x, 2), round(y, 2)) for x, y in cents], "inertia", round(inertia, 2))
def farthest_first(k):
chosen = [points[0]]
while len(chosen) < k:
chosen.append(max(points, key=lambda p: min(dist2(p, c) for c in chosen)))
return chosen
start_c = farthest_first(3)
cents, inertia = kmeans(start_c)
print("start C", start_c, "->", "inertia", round(inertia, 2))
Run with Python 3, it prints three lines, shown below. The twelve points are invented by us: three tight groups of four.
Starts A and C find the three obvious groups, with inertia 6.0. Start B stops at inertia 121.33.
What happened with start B
Start B places two centres inside the same bottom-left group and one centre between the other two groups. The two close centres split the bottom-left group between them, and the third centre sits at (7, 5), the average of eight points that belong to two different groups.
Nothing moves after that: every point is already nearest to its current centre, so the algorithm stops. It has converged, correctly by its own rules, to a poor answer. By our own division, its inertia is about twenty times that of the good solution.
What the documentation says
The scikit-learn guide is direct about it: "Given enough time, K-means will always converge, however this may be to a local minimum. This is highly dependent on the initialization of the centroids." Its remedies are to run the computation several times from different starts, and to use the k-means++ initialisation, which spreads the first centres apart.
Start C uses a simpler cousin of that idea, which we call farthest-first: take the first point, then repeatedly add the point farthest from every centre chosen so far. It picks (1, 1), (6, 9) and (9, 1), one per group, and finds the good answer.
Predict. Before running, sketch the points and the three starts on paper. Which start do you expect to fail, and where will its centres end up?
Choose k. Run k-means with k from 1 to 5 using farthest-first starts and print the inertia each time. Why does inertia always fall as k grows, and why does that make it a poor way to choose k on its own?
Break farthest-first. Add one point far away from everything, such as (20, 20). What does farthest-first choose now, and what does that tell you about outliers?
Compare with the library. Run scikit-learn's KMeans on the same points with n_init=1 and different random seeds, then with k-means++. Report how often each finds inertia 6.0.
The point is not k-means itself. It is the habit of distrusting a single run: of asking whether a result depends on a random start, a seed or an ordering, and checking. That habit applies to training a neural network, tuning a language model prompt or reading a published result, and it is the habit we most want university students to leave with.
English by age
How good is English among young adults in Peru?
It improves sharply through the early twenties. In EF's English Proficiency Index 2025, Peruvians aged 18 to 20 score in the "low" band, while those aged 26 to 30 reach "high".
EF EPI 2025, Peru, score by age group
Age group
EF EPI 2025 score
EF band
18 to 20
469
Low
21 to 25
540
Moderate
26 to 30
559
High
31 to 40
537
Moderate
41 and over
509
Moderate
Source: EF Education First, EF English Proficiency Index 2025, Peru country page, age-group scores and bands as published; Peru overall scores 519, "Moderate". Read on 9 November 2026. EF tests adults who choose to take its online test, so these are not university statistics.
Why it matters for AI
Most AI research papers, library documentation, model cards and courses are written in English, and so are error messages. The gap between a score of 469 and one of 540 is roughly the gap between needing a translator for a paper and reading it directly.
Taking an AI course in English while at university means the two skills grow together: the vocabulary of the subject (inertia, centroid, overfitting, embedding) becomes ordinary words, used in class every week.
How our classes handle it
Mentors speak clearly and at a steady pace, type key terms and formulas in the chat, and stop for questions. Students present their projects in short spoken summaries and written reports, both in English, with feedback on the explanation as well as the code.
We do not teach in Spanish. If you would rather read about our AI courses in Spanish first, see the curso de inteligencia artificial en inglés page; the classes it describes are the same live, English-taught classes.
Which course
Which AI course should a university student choose?
Start from what you can already do, not from what sounds most advanced. A student who skips the foundations usually returns to them within a month.
Choosing a first AI course (our recommendations)
If you
Start with
Then
Have never coded
Python Masterclass, or the AI and ML Masterclass from its first unit
Data analysis, then generative AI
Code a little, any degree
AI and ML Masterclass
Generative AI, or Search, Logic and Agents
Study CS or systems engineering
Search, Logic and Agents, beside DSA
Generative AI and agents
Want maths depth
Maths for Data Science
AI and ML Masterclass
Want to build with AI coding agents
Codex and Claude Code for professionals (18+)
Generative AI
Our own recommendations, adjusted for each student after the free first class.
Age and tools
Most undergraduates are 18 or over, so adult tools are open to them under the vendors' terms; Anthropic's Claude, for example, requires users to be at least 18. Some students start university at 16 or 17. For them we follow our teen rules: no Claude, and ChatGPT only with a parent's permission, as OpenAI's terms require under 18.
In every case, students first solve a problem without AI help, then use tools to check and extend their work, never to replace their understanding.
What a course will not do
It will not replace your degree, award Peruvian credits or guarantee a job. Students who finish a course receive our own certificate of completion. What employers and graduate programmes usually look at is the work: projects you can explain, tested properly, with code they can read.
A weather forecasting site with live conditions for any location.
by Krish
Fees
What the AI courses cost
You pay in US dollars with a card. The conversion to soles happens at your bank, on the day of each payment and at the bank's rate, so a statement can differ slightly from the previous month's.
Google reviews from students and families, quoted exactly.
★★★★★
"The one step solution for my son. Modern Age Coders make learning coding so simple that kids love it. The teachers explain complex concepts clearly with practical exercises and interactive content."
Ria Mukherjee
Parent
★★★★★
"Modern Age Coders has been a game-changer for me. I struggled to grasp IT concepts and coding before joining, but their classes transformed everything. I can now confidently write complex programs with ease."
Samriddha Mondal
Student
★★★★★
"One of the most wonderful education centres out there. Education is not limited to school syllabus but focuses on skill development."
Vansh Agarwal
Student
★★★★★
"My child Dhairya is really enjoying the Modern Age Coders classes. This is his first online class and he eagerly looks forward to it. I can already see his improvement, and the teachers are very cooperative."
Sonam Oswal
Parent of Dhairya
★★★★★
"Modern Age Coders have wonderful teachers who teach in a clear, easy and practical way. The teacher boosts students' confidence and inspires them to learn without hesitation."
Sonu Goyal
Parent
★★★★★
"I highly recommend this computer coding class! The teachers are incredibly knowledgeable and passionate about coding."
Ritu Kedia
Parent
Students ask
AI course for university students: questions
Are the classes in Spanish?
No. All classes are live online sessions taught in English, with no Spanish-language option. Mentors explain at a measured pace, write key terms and formulas in the chat and adjust to your accent, so a student with intermediate English can follow and improve as the course goes on.
Do I need to study computer science to take an AI course?
No. Students from economics, engineering, business, health, law and the sciences take our courses. If you have never coded, we start with Python; the maths needed for machine learning is taught as it comes up.
How much maths do I need?
Secondary school algebra is enough to begin. Statistics, probability and linear algebra are introduced as each model needs them. Students who want more depth can take our Maths for Data Science course alongside.
Does Peru require universities to teach AI?
DS 115-2025-PCM, published 9 September 2025, promotes the progressive addition of AI content to programmes such as systems engineering, software engineering, computer science and mathematics. It promotes rather than sets a fixed date for every programme; this is context, not legal advice.
When are classes held?
Any day, any time, flexible: choose a batch, or for a very personalised plan choose 1-on-1 at the hour that suits your family. Peru stays on UTC minus 5 all year and India on UTC plus 5:30, so 09:00 in Lima is 19:30 in India.
Can I use Claude or ChatGPT in the course?
If you are 18 or over, yes, within each vendor's terms, after you can solve the problem yourself. Students under 18 follow our teen rules: Claude is not used, because Anthropic's terms require users to be 18, and ChatGPT only with a parent's permission.
Will this help me get a job?
We make no job or salary promises. What the course gives you is skill and evidence: projects you built, tested on unseen data and can explain in English. Those are what you show an employer or a graduate programme.
Is there a certificate?
Yes, our own certificate of completion. It is not a university credit or an official Peruvian qualification, and we are not connected with SUNEDU or any university in Peru.
What computer do I need?
Any ordinary laptop with a stable connection and a headset. The early units run plain Python; later units use free cloud notebooks where heavier computation is needed, so a gaming graphics card is never required.
Can I pause during exams?
Yes. Tell your mentor your university calendar. One-to-one students can move classes around exams; group students can pause a month and rejoin a suitable group.
The trial is a working session: you write a few lines of Python, fit a small model and test it, while the mentor gauges your coding, maths and English. You leave with a recommended course and a plan that fits your semester.
Live video classes in English, taught by mentors in India. For a fast reply, write to us on WhatsApp in English.
Want a class sooner? Book a priority demo
A priority demo is a full class of about 45 to 60 minutes. After the class the mentor sends a written skill report and a personal learning roadmap on WhatsApp. If you enrol, the demo fee is adjusted against your first month's fee. Share your demo Order ID with us on WhatsApp and we take it off.
