About Practical AI Projects
The fastest way to learn AI is to build with it. This course applies your knowledge
through hands-on projects, from prediction systems that forecast
prices and customer behaviour, to computer vision applications that
recognise and locate objects in images.
You'll then build NLP solutions that understand and summarize text,
and recommendation engines that suggest movies and products using
collaborative filtering and embeddings.
Finally, you'll take your work into the real world with AI deployments:
serving a model through an API, packaging it with Docker, and monitoring it in
production.
Each project follows the same workflow: define the problem, prepare the data, build
and evaluate a model, and present the result. A working knowledge of Python and the
core ideas from Introduction to AI will help, and every project gives
you something concrete to add to your portfolio.
Content
10 projects
01
House Price Predictor
Build a regression model that predicts property prices from real housing data: explore the dataset, engineer features, and compare linear models with tree-based ensembles.
Prediction
Regression
Feature Engineering
Scikit-learn
02
Customer Churn Prediction
Predict which customers are likely to leave using classification. Handle imbalanced data, tune models, and judge them with precision, recall, and ROC curves.
Prediction
Classification
Imbalanced Data
Model Evaluation
03
Image Classifier
Train a convolutional neural network to classify images, then boost accuracy with data augmentation and transfer learning from a pretrained model.
Computer Vision
CNNs
PyTorch
Transfer Learning
04
Object Detection App
Detect and label multiple objects in photos and live video using a pretrained detector, bounding boxes, and confidence thresholds.
Computer Vision
Object Detection
YOLO
OpenCV
05
Sentiment Analysis
Classify product reviews as positive or negative. Start with a TF-IDF baseline, then fine-tune a transformer and compare the results.
NLP
Text Classification
TF-IDF
Transformers
06
Text Summarizer & Q&A Tool
Build a tool that condenses long articles and answers questions about them using pretrained language models from Hugging Face.
NLP
Summarization
Question Answering
Hugging Face
07
Movie Recommendation Engine
Recommend films using collaborative filtering and matrix factorization on user ratings, and handle the cold-start problem for new users.
Recommendation
Collaborative Filtering
Matrix Factorization
Cold Start
08
Semantic Product Recommender
Recommend similar products from their descriptions using text embeddings and vector similarity search.
Recommendation
Embeddings
Cosine Similarity
Vector Search
09
Model API with FastAPI & Docker
Wrap a trained model in a REST API with FastAPI, validate inputs, and package it in a Docker container ready to run anywhere.
Deployment
FastAPI
Docker
REST APIs
10
End-to-End AI Deployment
Take a model from notebook to production: build a reproducible pipeline, deploy to the cloud, and monitor predictions for drift.
Deployment
MLOps
Cloud Deployment
Monitoring
What You Can Do After This Course
By the end of the course you'll have shipped ten real AI projects spanning five
domains. Concretely, you'll be able to:
Build prediction systems — turn raw tabular data into regression and classification models you can evaluate and trust.
Create computer vision applications — train image classifiers and object detectors using CNNs and transfer learning.
Ship NLP solutions — classify, summarize, and answer questions over text with classic and transformer-based methods.
Design recommendation engines — combine collaborative filtering and embeddings to suggest relevant items to users.
Deploy real-world AI — serve models through APIs, containerize them, and monitor them in production.
Show proof of skill — ten finished projects across five AI domains, ready for your portfolio and GitHub.