Skip to main content
LearnMLAcademy
HANDS-ON PROJECT HANDBOOKS

Build projects you have actually heard of

These are not short demo notebooks. Each handbook is designed to take you from an empty folder to a working application with exact tools, commands, files, checks, troubleshooting and a final result you can show.

12
recognizable projects
Step by step
open → click → run → check
Real tools
clearly listed before you start
Project naming note: familiar product names such as Netflix, ChatGPT, Amazon or Perplexity are used only to make the learning goal immediately recognizable. LearnMLAcademy is not affiliated with those companies.
PROJECT 1Beginner4-6 hoursHANDBOOK V1 READY

Can You Survive the Titanic? Build a Machine Learning Predictor

A classic hands-on classification project rebuilt as a complete beginner-friendly handbook from raw CSV to a working prediction app.

What the student builds

A web app that takes passenger details and predicts survival probability while comparing multiple classification models.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyMatplotlibSeabornScikit-learnLogistic RegressionRandom ForestGradioGitGitHub
Chapters covered

Python · EDA · Missing Data · Encoding · Scaling · Classification · Model Evaluation · Cross-Validation · Feature Engineering · Deployment

Start the handbook
PROJECT 2Beginner5-7 hoursBUILDING NEXT

What Is This House Really Worth? Build a House Price Predictor

Build a real-estate-style price estimator and compare linear, regularized and ensemble regression models.

What the student builds

A house-price prediction application that turns raw property features into an estimated selling price.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyMatplotlibSeabornScikit-learnLinear RegressionRidgeLassoRandom ForestXGBoostGradio
Chapters covered

Regression · Loss Functions · Regularization · Feature Engineering · Scaling · Pipelines · Ensembles · Evaluation · Deployment

Full handbook will follow the Titanic template
PROJECT 3Intermediate5-7 hoursBUILDING NEXT

Can AI Catch a Stolen Credit Card Transaction?

Learn how real fraud problems differ from ordinary classification because fraudulent transactions are rare.

What the student builds

A fraud scoring system that ranks suspicious transactions and lets you tune the decision threshold.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyScikit-learnimbalanced-learnSMOTELogistic RegressionRandom ForestXGBoostMatplotlib
Chapters covered

Classification · Imbalanced Data · Precision · Recall · F1 · ROC-AUC · PR-AUC · Threshold Tuning · Ensembles · Anomaly Detection

Full handbook will follow the Titanic template
PROJECT 4Beginner4-6 hoursBUILDING NEXT

How Amazon Knows What Kind of Customer You Are

Turn shopping behaviour into customer groups and learn why clustering can reveal patterns without labels.

What the student builds

A customer-segmentation dashboard that groups shoppers by behaviour and explains what each segment represents.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyMatplotlibSeabornScikit-learnK-MeansDBSCANHierarchical ClusteringPCA
Chapters covered

Unsupervised Learning · K-Means · DBSCAN · Hierarchical Clustering · PCA · Scaling · EDA · Visualization · Feature Engineering

Full handbook will follow the Titanic template
PROJECT 5Intermediate5-7 hoursBUILDING NEXT

Can We Predict Tomorrow's Sales? Build a Retail Forecasting System

Build a forecasting workflow that learns trend and seasonality from historical store sales.

What the student builds

A forecasting application that predicts future sales and compares naive, statistical and ML-based forecasts.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyMatplotlibStatsmodelsARIMAExponential SmoothingScikit-learnXGBoost
Chapters covered

Time Series · Trend · Seasonality · Moving Averages · Exponential Smoothing · ARIMA · Lag Features · Forecast Evaluation

Full handbook will follow the Titanic template
PROJECT 6Intermediate5-7 hoursBUILDING NEXT

Build Your Own Netflix-Style Movie Recommendation System

Build a recommendation engine that finds movies a user is likely to enjoy from ratings and movie features.

What the student builds

An interactive movie recommender with similarity search, collaborative signals and a simple recommendation interface.

Tools you will use
PythonVS CodeJupyterLabPandasNumPyScikit-learnCosine SimilarityMovieLensEmbeddingsGradio
Chapters covered

Similarity · Feature Engineering · Unsupervised Learning · Embeddings · Recommendation Systems · Evaluation · Deployment

Full handbook will follow the Titanic template
PROJECT 7Intermediate6-8 hoursBUILDING NEXT

Can AI Detect a Real Disaster Tweet?

Move from classic text features to modern Transformers while solving a recognizable NLP classification problem.

What the student builds

A text classifier that decides whether a tweet describes a real disaster and explains model confidence.

Tools you will use
PythonVS CodeJupyterLabPandasRegexNLTKScikit-learnTF-IDFNaive BayesLogistic RegressionHugging Face TransformersPyTorch
Chapters covered

NLP · Text Cleaning · TF-IDF · Naive Bayes · Embeddings · Transformers · Fine-Tuning · Evaluation

Full handbook will follow the Titanic template
PROJECT 8Intermediate6-8 hoursBUILDING NEXT

Teach AI to Read Handwritten Numbers - Build Your Own Digit Recognizer

Train a neural network and CNN on handwritten digits, then turn the model into a small image prediction app.

What the student builds

An app where a user uploads or draws a digit and the model predicts which number it sees.

Tools you will use
PythonVS CodeJupyterLabNumPyMatplotlibPyTorchTorchvisionNeural NetworksCNNData AugmentationGradio
Chapters covered

Deep Learning · Tensors · Activations · Loss · Backpropagation · Optimizers · CNN · Regularization · Augmentation · Deployment

Full handbook will follow the Titanic template
PROJECT 9Intermediate5-7 hoursBUILDING NEXT

Build Your Own ChatGPT-Style AI Content Creator

Build a practical GenAI application that creates, rewrites and summarizes content with structured outputs and safety checks.

What the student builds

A content-generation studio with multiple writing modes, reusable prompts, structured outputs, history and export.

Tools you will use
PythonVS CodeLLM APIPrompt EngineeringPydanticJSONGradioEnvironment VariablesGitGitHub
Chapters covered

Generative AI · Prompt Engineering · Structured Output · Model Parameters · Evaluation · Responsible AI · Deployment

Full handbook will follow the Titanic template
PROJECT 10Intermediate6-8 hoursBUILDING NEXT

Chat With Your PDFs - Build Your Own RAG Assistant

Turn PDF documents into a searchable knowledge base and answer questions with retrieved evidence and citations.

What the student builds

A document Q&A app that uploads PDFs, chunks text, creates embeddings, searches a vector index and answers with sources.

Tools you will use
PythonVS CodePyPDFSentence TransformersFAISSVector SearchLLM APIRAGGradioGitHub
Chapters covered

Tokenization · Embeddings · Semantic Search · Vector Databases · RAG · Chunking · Reranking · Evaluation · Hallucinations · LLMOps

Full handbook will follow the Titanic template
PROJECT 11Advanced8-10 hoursBUILDING NEXT

Build Your Own Perplexity-Style AI Research Assistant

Build a tool-using agent that searches, reads, keeps notes, verifies evidence and writes a cited research answer.

What the student builds

A research agent that plans a question, gathers sources, stores evidence, verifies claims and returns a cited report.

Tools you will use
PythonVS CodeLLM APIWeb Search APITool CallingReActAgent MemoryState GraphsGradioGitHub
Chapters covered

Agentic AI · Tool Calling · ReAct · Planning · Context Engineering · Memory · State Machines · Agentic RAG · Safety · Evaluation

Full handbook will follow the Titanic template
PROJECT 12Advanced8-12 hoursBUILDING NEXT

Take an AI Model From Laptop to Production

Turn a notebook model into an API-backed production service with testing, containers, monitoring and CI/CD.

What the student builds

A production-ready ML service with an API, Docker image, automated tests, deployment workflow, monitoring and drift checks.

Tools you will use
PythonVS CodeFastAPIUvicornDockerPytestGitGitHubGitHub ActionsMLflowEvidentlyREST API
Chapters covered

MLOps · Serving · Batch vs Online Inference · Testing · CI/CD · Model Registry · Monitoring · Drift · Reliability · System Design

Full handbook will follow the Titanic template

The handbook standard

Every finished project will include environment setup, exact folder structure, commands to run, complete code, expected outputs, checkpoints, common errors, final application, deployment guidance, interview questions and real screenshots captured from the actual tools used.