Code the Future
What Computer Science and Artificial Intelligence really are, and how they are changing engineering, science, business and the humanities.
Engineering · Science · Business · Humanities — this room is for everyone
Your host todayHossem Eddine Kentache
Computer Science lecturer at NCUK · software engineer · co-founder & CTO
- Co-founded a fintech app: payments, virtual cards and live video shopping
- Built an Algerian search engine and an AI assistant
- Backend engineer for an international tech platform (France)
- Speaker at ICT Africa Summit 2026 · MSc & BSc, USTHB Algiers
I didn't learn all this to stay in front of a screen. I learned it to build things people use.
Which path are you on?
Click a card to add one · right-click to remove
The journey
12 live demos · you will flip real bits, build an Instagram post, train an AI and land a rocket.
What is Computer Science, really?
Spoiler: it's not about computers. Like astronomy isn't about telescopes.
Computer Science is the science of solving problems with step-by-step thinking
"It's for geniuses who love maths and sit alone hacking."
Hollywood hackers, green text, hoodies.
It's teamwork, creativity and logic, applied to any problem.
Doctors, bankers, architects, linguists and artists all use it. Writing code is just the tool.
Every app on earth uses these 4 ideas
Decomposition
Break a huge problem into small ones. "Build Uber" → map, payment, matching drivers.
Algorithms
Exact steps to solve each piece. A recipe a machine can follow, billions of times a second.
Data
Everything becomes numbers: photos, music, your voice, your DNA, a poem.
Abstraction
Hide the details. You drive a car without knowing how the engine burns fuel.
From electricity to Instagram, in 8 layers
The journey of a like
From a spark of electricity to a photo on Instagram. Lights down, sound up.
Esc to exit · M to mute
A computer only knows one thing: is the current ON or OFF?
⚡ ON = 1 · OFF = 0
That's all a computer "sees". No letters, no photos, no music: just current flowing, or not.
🔬 The transistor
A switch with no moving parts, controlled by electricity itself. Smaller than a virus: a modern phone chip has about 20 billion of them.
⏱️ Insanely fast
Each one can switch on and off billions of times every second.
8 switches = 1 byte = 256 possibilities
Same 8 switches, three meanings. The program decides how to read them. Try making 65 (the letter A) or 255 (all on).
Type your name. See how a computer really sees it.
Connect switches cleverly and they start to decide
Click the inputs. AND: the door opens only with your card AND the right PIN. OR: the alarm rings if the door OR the window opens. Every decision a computer makes is built from these.
Gates that can add: the heart of every calculator
Click the bits of A and B. The answer is computed only with AND, OR and XOR gates, exactly like inside a chip.
A phone processor does this billions of times a second. Multiplying = adding many times. Photos, games and AI = an ocean of additions.
The CPU: a tireless worker that repeats 3 steps, forever
get the next instruction from memory
understand it: "add these two numbers"
do it with the logic gates
= 3 billion cycles every second
fast memory: what the phone is using right now
your photos and apps, kept even when it's off
Humans can't write in 0s and 1s, so we invented languages
10111000 00000001 00000000 00000101 00000010 00000000 11001101 00100001
MOV AX, 1 ADD AX, 2 INT 21h
int total = 1 + 2;
printf("%d", total);total = 1 + 2 print(total)
A compiler or interpreter translates the nice language down to 0s and 1s, like a translator between you and the chip.
Your screen is millions of tiny lights, each one made of 3 numbers
Click the heart to zoom in and see the numbers inside every pixel.
Write code → see an Instagram post appear
This is real HTML + CSS, the languages of every web page. Change the text, the colour or the emoji and watch.
You tap "like". Your friend in London sees it in less than a second.
Algeria is connected to Europe by several undersea fibre-optic cables across the Mediterranean. Your like literally travels as light under the sea.
⚡ → 0️⃣1️⃣ → 🚦 → ➕ → 🧠 → ✍️ → 🖼️ → 📱
Instagram is not magic. It's billions of tiny switches, organised by people who learned to think in layers.
And every one of those layers was built by someone who was once a student, like you.
From one woman's notes to ChatGPT in 180 years
Notice: two of the founders of computing were women, and one was a poet's daughter (Ada was Lord Byron's daughter).
CS is not one job. It's a universe of them.
AI & Data
Machines that learn from examples.
Cybersecurity
Protect banks, hospitals, countries.
Software & Apps
Web, mobile, games, the apps you use daily.
Robotics
Code that moves in the real world.
Networks & Cloud
How 5 billion people connect.
UX & Design
Making technology human.
Quantum computing
The next frontier of computing.
Graphics & VR
Films, games, simulations, like this 3D world.
What is AI, really?
Not a robot that wants to take over the world. A very powerful pattern-finding machine.
Normal programs follow rules. AI learns the rules from examples.
Rules + data → answer
"If the email contains 'you won a prize', mark it as spam." A human writes every rule. It breaks when spammers change words.
Data + answers → rules
Show it 1 million emails labelled spam / not spam. It discovers the patterns itself, even ones no human noticed.
AI ⊃ Machine Learning ⊃ Deep Learning (neural networks) ⊃ Generative AI (ChatGPT, image generators)
Teach an AI to separate red from blue
Click to add red points · Shift+click (or switch) to add blue points. Then press Train and watch the AI learn the border.
A tiny neural network: 2 inputs → 8 neurons → 8 neurons → 1 output. It adjusts its 105 numbers ("weights") again and again to make fewer mistakes. That's all "learning" is.
It predicts the next word. Again. And again. And again.
1 · Read
It's trained on a huge part of the internet, books and code: trillions of words.
2 · Predict
Given "The capital of Algeria is", it learns that "Algiers" is the most likely next word.
3 · Refine
Humans rate its answers, so it learns to be helpful and polite (RLHF).
It doesn't "know" things like you do. It's an extremely good guesser, which is why it can sound sure and still be wrong.
Your own mini-ChatGPT, trained on 40 sentences
This model only counts which word usually comes after another word in its training text. Click a word to add it, or let it write.
Real LLMs do the same thing, but they look at thousands of previous words, not just one, and they learned from trillions of words, not 40 sentences.
AI is a superpower… if you stay the boss
Great at
Explaining, summarising, brainstorming, translating, drafting, writing code, finding patterns in data.
Bad at
Being always right. It can invent facts, sources and quotes ("hallucinations"). It can repeat biases in its data.
At NCUK
Use it to learn, not to cheat: your work must be your own, and you must check what AI tells you.
The best prompt formula: Role + Task + Context + Format. "You are a biology tutor. Explain osmosis to a foundation student, with one everyday example, in 5 lines."
☕ Coffee & questions
Next: rockets, robots and engineering. Come back on time, there's a game!
Engineering, rebuilt by code
Today, the most powerful tool of a mechanical, civil or electrical engineer is not a wrench. It's software.
A 40-metre rocket that lands by itself
Reusable rockets
Since December 2015, SpaceX Falcon 9 boosters fly back and land upright. Some have flown more than 20 times.
Software computes the landing path in real time, many times per second.
Guidance algorithms
Optimisation algorithms (convex optimisation) find the fuel-saving path to the landing pad, while wind pushes the rocket.
Landing on Mars
In 2021 NASA's Perseverance compared camera images to maps during descent and chose a safe spot by itself: Terrain-Relative Navigation.
Land the rocket. Softly. Then let the autopilot try.
Hold Space (or the button) to fire the engine. ← → to steer. Land at less than 2 m/s on the pad.
The autopilot is a PD controller, 6 lines of code engineers use everywhere: drones, cruise control, robot arms, even your fridge.
Crash a car 1,000 times before building it once
Computational fluid dynamics
Planes, F1 cars and wind turbines are tested in virtual wind tunnels: millions of tiny air cells solved by computers.
Civil engineering
Finite element analysis simulates how bridges and towers bend in earthquakes, useful in a seismic country like Algeria.
Digital twins
A live virtual copy of a real factory, engine or city, fed by sensors. Test changes on the copy, not on the real thing.
Industry 4.0: machines that see, predict and adapt
Robots
Factory arms, warehouse robots, surgical robots and walking robots like Boston Dynamics' Atlas.
Predictive maintenance
Jet engine makers stream sensor data from engines in flight and fix parts before they fail.
Self-driving
Robotaxis (e.g. Waymo) carry paying passengers with no driver in several US cities.
Smart energy
AI balances power grids with solar and wind, a big topic for Algeria's solar future.
Engineer + code = unstoppable
🐍 Python
Data, simulation, automation, AI. The #1 language in science and engineering.
📐 MATLAB / Simulink
Control systems, signals, vehicles. Used in industry and universities.
🔌 Arduino & Raspberry Pi
Build real robots and sensors at home for a few thousand dinars.
Weekend project idea: a plant that waters itself (soil sensor + Arduino + 15 lines of code).
Science at the speed of AI
Problems that took 50 years are being solved in months.
AlphaFold: a 50-year biology puzzle, solved
The protein folding problem
Proteins are chains that fold into 3D shapes, and the shape decides what they do in your body. Finding one shape used to take scientists years of lab work.
AlphaFold (Google DeepMind) predicted the structures of about 200 million proteins: nearly every known protein.
Nobel Prize in Chemistry
Demis Hassabis and John Jumper (AlphaFold), with David Baker (protein design). AI is now a Nobel-level science tool.
Your next doctor will have an AI assistant
Medical imaging
AI helps radiologists spot cancers in mammograms and scans, and catches things tired eyes miss.
Eye disease
AI screens retina photos for diabetic retinopathy, a big help where specialists are rare.
New drugs
In 2020 MIT researchers used deep learning to discover halicin, a new antibiotic against resistant bacteria.
Wearables
Smartwatches detect irregular heartbeats (atrial fibrillation) and alert the user to see a doctor.
From molecules to black holes
Chemistry & materials
AI searches millions of virtual materials for better batteries, solar cells and catalysts before anyone mixes a chemical.
Weather & climate
Google DeepMind's GraphCast (2023) produces 10-day forecasts in under a minute, with accuracy that rivals traditional supercomputer models.
The first black hole photo
In 2019 algorithms combined data from 8 telescopes across the planet into one Earth-sized virtual telescope: about 5 petabytes of data.
Your DNA is 3 billion letters of code
Bioinformatics
Biology + computer science: read genomes, find disease genes, track viruses. During COVID-19, scientists shared and analysed virus genomes worldwide within days.
Personalised medicine
Treatments chosen for your genes. Cancer therapies are already matched to the tumour's DNA.
Science students: Python + statistics are now as important as the lab coat.
🥤 Stretch break
Next: money, markets and the apps that know what you want.
Business runs on data
"Data is the new oil" — but only if you know how to refine it.
Your money is now software
Fraud detection
Every card payment is scored by AI in milliseconds: normal for you, or suspicious?
Mobile banking
Pay bills and transfer money from your phone. In Algeria: BaridiMob, CIB and Edahabia cards.
Credit scoring
Algorithms decide who gets a loan, which raises big questions about fairness.
Algorithmic trading
Most trades on big stock markets are placed by computer programs.
Pick 3 things you love. The recommender does the rest.
Choose 3 items…
The system compares you with 8 imaginary users and recommends what similar people liked: "collaborative filtering", the idea behind Netflix, YouTube, Amazon and TikTok.
Every department now has an AI co-worker
Marketing
Ads targeted to the right people, A/B tests, AI-generated visuals and slogans.
Sales
CRM tools predict which customer will buy next and when to call them.
Supply chain
Forecast demand, optimise delivery routes, keep stock low but never empty.
Customer service
Chatbots answer simple questions 24/7; humans handle the hard ones.
Turn numbers into decisions
Analyse
Excel → SQL → Python. What sold, where, when, and why?
Predict
Next month's sales, which customers will leave, the best price.
Tell the story
Dashboards (Power BI, Tableau) that a CEO understands in 10 seconds.
Algerian start-ups like Yassir (ride-hailing & delivery super-app) are built on exactly this: data, algorithms and apps.
The humanities are going digital
The most human questions — language, history, law, mind, art — now have powerful new tools.
Reading a scroll burned by a volcano 2,000 years ago
The Vesuvius Challenge
The Herculaneum scrolls were carbonised by the eruption of Vesuvius in 79 AD: too fragile to open. Scientists scanned them with X-rays, and in 2023 a student used AI to read the first word inside a closed scroll: "purple" (πορφύρας).
By 2024, whole passages of an ancient philosophy text had been read.
Lost cities
LiDAR (laser scans from planes) plus algorithms revealed tens of thousands of hidden Maya structures under the Guatemalan jungle.
Words are data too
Languages
Machine translation, speech recognition, and AI for low-resource languages like Darija and Tamazight. Linguists are needed!
Law (legaltech)
AI searches thousands of laws and contracts in seconds; lawyers decide what's fair.
Journalism
The Panama Papers (2016): 11.5 million leaked documents, investigated with data tools by journalists in 80 countries.
Psychology
Analyse language to study emotions and mental health; design apps that are good for us, not addictive.
Paste any text: a poem, a speech, a review
Mood
Most used words
Who decides what AI should do? Humanities.
Creative industries
Games, films, music and design use AI tools, but taste, story and meaning come from people.
AI ethics & policy
Bias, privacy, copyright, deepfakes, jobs. Governments (e.g. the EU AI Act) need people who understand both humans and tech.
Philosophy is back
What is intelligence? Consciousness? Responsibility? These are now practical engineering questions.
Whatever you love, there's a "-tech" for it
You don't need to become a programmer. You need to become a "bilingual" professional.
Engineering
Python + MATLAB, Arduino projects, CAD & simulation. Careers: robotics, aerospace, energy, automotive.
Science
Python + statistics, bioinformatics. Careers: medical AI, drug discovery, data scientist, health tech.
Business
Excel → SQL → Power BI → Python. Careers: data analyst, fintech, digital marketing, founder.
Humanities
AI literacy, prompting, text analysis, UX. Careers: AI ethics, legaltech, UX research, digital journalism.
Free to start: CS50 (Harvard, free online) · Google's free AI courses · Kaggle Learn · freeCodeCamp · Khan Academy
This week, use AI or code to solve one real problem in your field.
A summary of a hard lesson. A spreadsheet that does your budget. A chart of your sport stats. A translation tool for your grandmother. Small is perfect.
"AI won't replace you. A person who uses AI well might." Be that person.
Questions?
Hossem Eddine Kentache · Computer Science, NCUK Algeria
hossemeddinekentache@ncukalgeria.com