100 Days of Mastering AI
Learn Python by building AI systems, automations, and practical projects.
Computers π»π§ β‘
Understand how computers follow instructions and why this matters for AI builders.
Variables π¦πβ‘
Variables store information for your program.
Data Types πΊπ°π‘οΈπ‘
Different data needs different containers. String (text), Integer (whole numbers), Float (decimals), Boolean (True/False). Choose the right type or code breaks.
User Input π€β¨οΈπ¬
Programs become interactive when they listen to users. Learn input() to make your code talk to humans and respond.
Conditions β‘π€π¦
Code can make decisions using if statements. Programs evaluate TRUE/FALSE conditions and choose what to do. This is how apps respond differently to different situations.
Loops πβ‘βΎοΈ
Loops let computers repeat tasks automatically. No more copy-pasting code. Write it once, run it 1000 times.
Mini Project π―π»π§©
Combine everything from Week 1 into one real program. Variables + Input + Conditions + Loops working together.
Functions πβ‘π§
Write code once, use it forever. Functions are reusable machines that do your work on command!
Inputs & Outputs π₯π€βοΈ
Make functions smarter by sending data in and getting results back. This is how functions become truly powerful!
Lists ππ¦π
One variable that stores many values. Like a train with many compartments!
List Operations πβοΈβ‘
Lists can grow, shrink and change. Add, remove and update items just like a shopping cart!
List Loops πππ€
Loop through every item in a list automatically. No item gets skipped!
Dictionaries ππποΈ
Store data as key-value pairs. Like a contact book β search by name, get their number instantly!
Strings π¬π€βοΈ
Strings are how programs handle text. Every chat, search and message in every app is a string!
String Operations βοΈππ§±
Transform and combine text like LEGO blocks. Build bigger messages from smaller pieces!
Tuples ππ¦π‘οΈ
Tuples store data that should never change. Like a sealed locker β you can read it but never edit it!
Sets π―πβ
Sets automatically remove duplicates. Only unique values survive β like a VIP guest list with no repeats!
Nested Data π¦ππ
Data inside data! Like a house with rooms with cupboards β access anything layer by layer!
File Handling ππΎπ
Save data permanently so it survives after your program closes. Like writing in a notebook you can read anytime!
Error Handling π‘οΈβ οΈπ§
Stop your program from crashing! Handle problems gracefully so the app keeps running no matter what.
Modules & Imports π§°π₯βοΈ
Use tools built by others! Import powerful ready-made code instead of writing everything from scratch.
Packages & Pip π¦ππ οΈ
Install new superpowers! Use pip to add powerful packages and capabilities to your Python projects.
Virtual Envs π ππ¦
One project, one private room! Keep packages isolated so projects never conflict with each other.
OOP π€π§±ποΈ
Model real-world things in code. A class is a blueprint, an object is the real thing built from it!
Classes & Objects ποΈππͺ
Classes define the idea. Objects bring it to life. One cookie cutter, unlimited cookies!
Constructors (**init**) π₯ππͺͺ
Give every object an identity the moment it's created. Like filling a registration form automatically at birth!
Inheritance π¨βπ©βπ§βπ¦ππ
Get everything for free! Child classes inherit features from parent classes and add their own powers.
Encapsulation ππ¦π‘οΈ
Lock your data, control access! Protect important information and allow access only through approved methods.
Polymorphism ππβ‘
One command, many responses! The same method can behave differently depending on the object using it.
Student System ππποΈ
Build a real software project using classes, inheritance, encapsulation and constructors together!
Internet Basics ππ¦π£οΈ
Understand the invisible road your data travels. Every app open = thousands of packets racing across the world!
HTTP ππ¨π₯οΈ
Understand how browsers and servers communicate. Every website visit is a request and response conversation!
APIs ππβ‘
Learn how APIs let your apps connect to maps, payments, weather, AI, and other powerful services without building everything from scratch.
JSON Data Format πππ
JSON is the universal language every system understands. Like a passport β same format, accepted everywhere!
API Requests π±ππ»
One line of Python. The entire internet at your fingertips. requests.get() connects your code to any server on the planet!
Public APIs ππβ‘
Real data. Zero cost. One line of Python. Public APIs are free libraries β walk in, grab your data, walk out!
API Auth πππ‘οΈ
API keys prove who you are. Like a membership card, they unlock access to powerful APIs while keeping everyone else out.
Backend vs Frontend βοΈπ»π
Two worlds, one product. Frontend is what you see. Backend is the kitchen working hard so you never have to think about it!
Databases ποΈπΎπ
Give your program permanent memory. Variables die when code stops. Databases live forever!
SQL Basics ποΈπβ‘
Give your program permanent memory. Variables die when code stops. Databases live forever!
SQL Queries ππβ‘
Ask the right question, get the right answer. SQL queries help you filter, sort and find exactly the data you need!
Filtering Data πβ π
Stop fetching everything. Use WHERE to fetch only what matters β like Swiggy's filter panel but for your database!
Sorting Data ππβ¬οΈβ¬οΈ
Every list becomes a leaderboard with ORDER BY. Sort highest first, lowest first β raw data becomes meaningful instantly!
SQL Joins πππ
Two tables. One shared key. Infinite answers. JOIN links separate tables together to give you the complete picture!
DB Design πποΈπ
Plan your tables like an architect plans rooms. Design first, code second β or you'll rebuild everything later!
Python + SQLite ποΈπβ‘
One file. Zero setup. Full database power. SQLite is Python's personal diary β open, write, read, close. No server needed!
CRUD Applications βπβοΈποΈ
Four operations. Every app ever built. Mastered. Create, Read, Update, Delete β Instagram, Zomato, Gmail β just CRUD!
Data Modeling πππ
Draw the boxes first. The code writes itself after. A data model is the skeleton β everything else is built on top of it!
Expense Tracker β Phase 2 Complete! ππ°ποΈ
49 days. Real data. Your first backend product. Build a working expense tracker that remembers every rupee β permanently!
System Design ποΈβ‘π
Day 50 β halfway to AI Engineer! Now you think like a systems engineer. Every piece planned before traffic arrives!
Data Science π¬ππ§
Turn raw numbers into decisions that matter. Data science is the bridge between drowning in data and finding wisdom in it!
NumPy β‘π’π
Learn how NumPy supercharges Python with lightning-fast arrays. Process entire datasets in one line β no loops, no pain.
Arrays & Vector Thinking β‘π’π
One operation. Every element. Zero loops needed. Vector thinking is how engineers move thousands of values at once β instantly!
Pandas Introduction πΌπβ‘
Excel for programmers. Infinite scale. Zero mouse clicks. Pandas turns any dataset into a conversation β and you're finally fluent!
DataFrames ππΌπ
Master Pandas DataFrames β filter rows, select columns, peek at data, and calculate stats. Ask precise questions, get exact answers every time.
Data Cleaning π₯¦π§Ήβ
Clean data is not a luxury β it's a requirement. Garbage in, garbage out. Clean first, analyze second. Never skip order!
Data Visualization ππ¨πΊοΈ
See the pattern. Tell the story. Change the decision. Charts make the invisible impossible to ignore!
Matplotlib Basics ππ¨βοΈ
Master the tools. Then let the data speak. Matplotlib is an artist's sketchbook β same tools, endless combinations!
Exploratory Data Analysis ππ π
Know your data before you trust your data. EDA is the house inspection β find every crack before it costs you later!
Data Dashboard β Phase 3 Complete! πππ
60 days. Three complete phases. Build a full analyst dashboard that cleans data, finds patterns and draws charts automatically!
What Is Machine Learning? π€π§ π
Stop writing rules. Start teaching patterns. ML is when you show examples and let the algorithm figure out the rules itself!
Features & Labels ππ’π―
What goes in. What comes out. The whole game. Features are the inputs the model sees. Labels are the answers it must predict!
Training vs Testing Data πβ π
Train on the past. Test on the unknown. Always. A perfect training score means nothing β only the test score tells the truth!
Regression Explained ππ π’
Give it inputs. It gives you a number. That's regression. Predict house prices, exam scores, temperatures β any continuous number!
Classification ποΈπ€β
Learn how machine learning classification predicts which category an input belongs to. Build your first classifier using scikit-learn and LogisticRegression.
Scikit-Learn π§π€π¦
Master the one universal workflow that powers every scikit-learn model β Import, Instantiate, fit(), predict(). One pattern. Every algorithm. Infinite models.
ML Model ποΈπ€π
Bring all your ML skills together into one complete pipeline β data, split, train, predict, and evaluate. Five lines of code. Your first real working model.
Evaluation ππ―π
Learn to evaluate your ML model using MAE and RΒ² on unseen test data. Training score means nothing β only the test score is the truth.
Overfitting ππβοΈ
Diagnose the two biggest ML problems β overfitting and underfitting. Learn to detect them using the train vs test RΒ² gap and build models that generalise perfectly.
Prediction App ππ€π
Phase 4 complete β combine every ML skill into one real working prediction system. Data, split, train, evaluate, predict. A full pipeline any data team would recognise.
Deep Learning π€π§ π
Learn how deep learning uses stacked layers to automatically find patterns in raw data β no rules written by hand β and build your first neural network with TensorFlow and Keras.
Neurons β‘π§ π
Understand how a single artificial neuron works β inputs multiplied by weights, plus a bias, checked against a threshold β and why this tiny decision unit powers every AI system in the world.
Neural Nets π§ πβ‘
Learn how a neural network stacks layers of neurons to transform raw data into answers β and why activation functions like ReLU are the secret that makes depth actually matter.
Forward Pass β‘οΈπ§ β‘
Learn exactly what happens inside a neural network when it makes a prediction β data flows one direction through four steps: multiply, add bias, activate, and pass forward β and why weights never change going forward.
Activations β‘π§ π―
Learn what activation functions do between layers β how ReLU kills negatives to keep learning alive, how Sigmoid squashes outputs into probabilities, and the golden rule: ReLU for hidden layers, match output activation to your task.
Loss Functions π―ππΉ
Discover how loss functions act as an AI model's coach by measuring exactly how wrong each prediction is, and learn the three key types β MSE, MAE, and BCE β so your model can improve with every shot.
Backpropagation πππ§
Understand how backpropagation works like a cricket coach rewinding match footage to fix every mistake layer by layer, and see exactly how neural networks adjust their weights to get smarter with every epoch.
TensorFlow & PyTorch π₯π§ β‘
Learn the difference between TensorFlow and PyTorch β the two most powerful deep learning frameworks β understand when to use each one, and build your first neural network in both.
Training Networks π²ππ
Learn how neural networks train epoch by epoch β just like learning to ride a bicycle β and understand why validation data is the only honest way to know if your model is truly getting better.
Image Classifier πΌοΈπ€β‘
Put every Deep Learning concept from Phase 5 together and build a real image classifier that reads handwritten digits with 98% accuracy β your first complete end-to-end neural network project.
AI Basics π€π§ πΊ
Zoom out and understand the full AI landscape β from Narrow AI that exists today to General and Super AI on the horizon β and see exactly where everything you have built so far fits inside the bigger picture.
NLP ππ’π€
Learn how machines read human language by converting raw text into numbers using tokenization and vectorization, and build your first text pipeline using CountVectorizer to transform sentences into data a model can actually understand.
Embeddings πΊοΈππ’
Learn how embeddings give every word a location in meaning space so that similar words sit close together, and discover how to measure word similarity and perform word arithmetic using cosine similarity.
Transformers βοΈβ‘π§
Discover how Transformers revolutionised AI by processing all words simultaneously using attention β unlike RNNs that forget early context β and use HuggingFace to run a full pretrained Transformer in just one line of code.
Attention π¦ππ§
Learn how the attention mechanism gives every word a spotlight score so the model knows exactly which other words to focus on for context β and see why "bank" means something completely different depending on whether "river" is in the sentence.
LLM Basics π€π¬β‘
Demystify exactly how ChatGPT generates responses token by token using the six-step pipeline β tokenize, embed, attention, predict, append, repeat β and understand why it completes patterns instead of retrieving facts from a database.
Prompt Engineering π―βοΈβ‘
Master the four-part prompt formula β Role, Context, Format, and Constraints β that transforms vague 20% outputs into precise 90% results, and see exactly why the prompt is the musician and the model is just the instrument.
Context Windows π§ πͺβ‘
Learn how AI models use a fixed-size context window as their working memory, why older messages get dropped when it fills up, and how to write smarter prompts that stay within limits.
Tokens π€π€β‘
Learn what tokens really are, how AI breaks your text into bite-sized chunks before reading it, and why counting tokens is the secret to controlling cost, speed, and context limits in every AI app you build.
AI Chatbot π€π¬π
Put every AI and LLM concept from Days 81-89 to work by building a real multi-turn AI chatbot powered by LLaMA 3 via Groq API β with memory, personality, token tracking, and a context window guard.
RAG πππ€
Learn how RAG stops AI from making things up by retrieving real, fresh context before generating answers β like giving a genius student the right textbook pages during an open book exam.
Vector DBs ποΈπ§ π
Learn how vector databases store text as coordinates in meaning space and find the nearest neighbors in milliseconds β the secret engine powering every serious AI system.
Semantic Search ππ§ β‘
Learn how semantic search understands the meaning behind your words β not just exact letters β using vectors and cosine similarity to return results that actually make sense.
AI Agents π€π οΈβ‘
Learn what makes an AI agent fundamentally different from a chatbot β agents perceive, reason, act, observe, and loop autonomously until the goal is complete, taking multiple real-world actions with zero hand-holding after the first instruction.
Agent Tools π οΈβ‘π€
Learn how AI agents use tools to take real-world actions β the LLM is the brain that decides which tool to call, when, and why, while each tool does exactly one precise job and chains its output to the next step.
Multi-Agents π€π€β‘
Learn how multi-agent systems work β one orchestrator breaks down a goal and assigns it to specialised agents, each handling only their domain, with structured outputs chaining between them to produce results better than any single agent alone.
AI Workflows βοΈπ€π
Learn how to build AI automation workflows that run end-to-end with zero human input β one trigger fires every step automatically, with error handling at each stage and human gates before any irreversible action.
Production AI ππ€ποΈ
Learn the difference between a demo that works on your laptop and a production AI system that serves real users while you sleep β covering the three pillars every production system needs: reliability, observability, and scalability.
AI Architecture ποΈπ€π
Learn how to design a lean production-ready AI startup stack β build the simplest thing that ships today, design scale doors for tomorrow, and never over-engineer before your first user. Every choice reversible. Ship first.
AI Platform π€ποΈπ
The final day β build a complete AI Agent Platform with four specialised agents, a tool registry, SQLite memory, and multi-agent orchestration. Every concept from Day 1 to Day 99 combined into one production-ready system built entirely by you.