Artificial Intelligence Engineering (AI Engineer)

On the AI Engineering course, you'll learn data analysis, machine learning models, boosting, clustering, recommender systems and the basics of deep learning through real projects.

UX/UI design Backend development Digital marketing

Course program

26 topic 90 lesson
01 Understanding the problem and initial checks on raw data 5 lessons
  1. 1
    Framing the business / research question
    • Clarifying the objectives
    • Defining the target variable
    • Success metrics
  2. 2
    Importing data sources
    • Flat files (CSV, Excel)
    • SQL queries
    • APIs
  3. 3
    A first look in pandas
    • head(), info(), describe(), data types, memory usage
  4. 4
    Data structure concepts
    • Rectangular and nested data
    • Tidy data principles
    • Indexes
  5. 5
    Quick data quality checks
    • Counting missing values
    • Duplicate rows
    • Detecting impossible values
02 Handling missing data and outliers 5 lessons
  1. 6
    Diagnosing missingness
    • MCAR, MAR, MNAR
    • Visual tools (missingno, heatmaps)
  2. 7
    Imputation strategies
    • Mean / median
    • Constant value
    • Flag columns
  3. 8
    Detecting outliers
    • Z-score
    • The IQR rule
    • Tukey fences
    • Domain-based limits
  4. 9
    Robust scaling
    • Winsorization
    • Log / Box-Cox transformations to reduce skewness
  5. 10
    Practical tips
    • When to drop and when to impute
    • Documenting every assumption
03 Exploring data characteristics 5 lessons
  1. 11
    A deep dive into descriptive statistics
    • Mean, median, mode, trimmed mean
    • MAD
    • Coefficient of variation (CV)
  2. 12
    Distribution analysis
    • Histograms, KDE, QQ plots
    • Detecting skewness and kurtosis
  3. 13
    Visual exploration of categorical data
    • Frequency tables
    • Bar charts
    • Pareto charts
  4. 14
    Relationships in numerical data
    • Scatter matrices
    • Correlation heatmaps
    • Pair plot diagnostics
  5. 15
    Initial observations on features
    • Valuable signals, redundancies and surprises for later feature engineering
04 Data transformation and feature preparation 4 lessons
  1. 16
    Encoding categorical variables
    • One-hot and ordinal encoding
    • Target and frequency encoding
    • The risks of high cardinality
  2. 17
    Scaling and normalization
    • StandardScaler, Min-Max, RobustScaler
    • Choosing based on the algorithm
  3. 18
    Feature creation basics
    • Date-time components, aggregations, ratios, interaction features
  4. 19
    Building pipelines in scikit-learn
    • ColumnTransformer
    • Custom transformers
    • Reproducible workflows
05 Lab: EDA and data cleaning 5 lessons
  1. 20
    Loading a real mixed-type dataset
  2. 21
    Systematic handling of missing values and outliers
  3. 22
    Creating distribution and relationship visualizations
  4. 23
    Building a preprocessing pipeline that outputs a model-ready table
  5. 24
    Presenting a short notebook summarizing the findings and quality metrics of the cleaned data
06 Machine learning: linear models I – simple and multiple regression 3 lessons
  1. 25
    Regression basics
    • Hypothesis and ordinary least squares (OLS)
    • Computing the cost function: MSE and the normal equation
  2. 26
    Multiple regression
    • Vectorized implementation
    • How feature scaling affects solvers
  3. 27
    Model evaluation
    • Metrics: MSE, RMSE, R2
    • Cross-validation overview
07 Linear models II – polynomial regression and interactions 3 lessons
  1. 28
    Polynomial regression
    • Expanding features with PolynomialFeatures
    • Choosing the polynomial degree
    • Effect on model capacity
  2. 29
    The bias–variance trade-off
    • Signs of underfitting and overfitting
    • The effect of model complexity
  3. 30
    Learning curves
    • Plotting training and validation error
    • Diagnosing capacity issues
08 Regularized regression: Ridge, Lasso, Elastic Net 4 lessons
  1. 31
    Why regularization?
    • Recap of the signs of overfitting
    • Shrinkage and coefficient paths
  2. 32
    Ridge regression (L2)
    • Cost function with λ |w|²
    • Effect on correlated features
    • Implementation with the closed-form solution and SGD
  3. 33
    Lasso regression (L1)
    • Sparse solutions and feature selection
    • Geometry: the diamond-shaped constraint region
    • When to prefer it over ridge
  4. 34
    Elastic Net
    • Combining L1 and L2 penalties
    • Tuning α and l1_ratio with grid search CV
09 Logistic regression 3 lessons
  1. 35
    Logistic regression
    • Sigmoid activation and probability estimates
    • Cross-entropy loss and gradient descent
  2. 36
    Naïve Bayes
    • Bayes' theorem with the feature independence assumption
    • Gaussian and Multinomial variants
  3. 37
    Model evaluation
    • ROC curve and AUC
    • Precision, recall and F1-score
10 Regression and classification lab 2 lessons
  1. 38
    Regression practice
    • Building linear and polynomial models
    • Experimenting with ridge and lasso regularization
  2. 39
    Classification practice
    • Training logistic and Naïve Bayes classifiers
    • Evaluating with ROC and precision/recall
11 K-nearest neighbors (KNN) 3 lessons
  1. 40
    KNN basics
    • Distance metrics and choosing k
    • Uniform and distance-weighted voting
    • The effect of feature scaling
  2. 41
    High-dimensional data
    • The curse of dimensionality
    • Nearest neighbor search strategies
    • Dimensionality reduction as preprocessing
  3. 42
    Model evaluation
    • Cross-validation for choosing k
    • The confusion matrix and classification report
12 Support vector machines (SVM) 3 lessons
  1. 43
    Maximum margin classifiers
    • Hard and soft margin formulations
    • Slack variables and the C parameter
  2. 44
    Kernel trick
    • Linear, polynomial and RBF kernels
    • Implicit feature mapping
  3. 45
    Model tuning
    • Balancing C and γ
    • Grid search and scaling requirements
13 SVM project 4 lessons
  1. 46
    Data preparation
    • Loading a two-class dataset (e.g. make_moons)
    • Train/test split and applying StandardScaler
  2. 47
    Linear SVM
    • Training a LinearSVC
    • Evaluating with accuracy, the confusion matrix and a classification report
    • Plotting the decision boundary
  3. 48
    Kernel SVM
    • Training SVC(kernel='rbf') and SVC(kernel='poly', degree=3) models
    • Evaluating and plotting each one
  4. 49
    Hyperparameter tuning
    • GridSearchCV over C ([0.1,1,10]) and gamma ([0.01,0.1,1])
    • Retraining the best model and reporting test accuracy
14 Decision trees 3 lessons
  1. 50
    Building a tree
    • Split criteria: Gini and entropy
    • Recursive binary splitting
  2. 51
    Tree regularization
    • Pre-pruning: max_depth, min_samples
    • Post-pruning methods
  3. 52
    Interpretation
    • Feature importance
    • Visualizing the tree
15 Random forest 3 lessons
  1. 53
    Bagging and variance reduction
    • Bootstrap sampling
    • Aggregating predictions
  2. 54
    The random subspace method
    • Random feature selection at each split
    • Decorrelated trees
  3. 55
    Out-of-bag evaluation
    • Estimating the OOB error
    • Feature importance based on OOB samples
16 Tree ensembles lab 3 lessons
  1. 56
    Decision tree practice
    • Training and visualizing a single tree
    • Experimenting with pruning parameters
  2. 57
    Random forest practice
    • Building an RF model and measuring the OOB error
    • Computing and plotting feature importance
  3. 58
    Hyperparameter tuning
    • Grid search over n_estimators and max_features
17 Boosting 5 lessons
  1. 59
    Boosting concepts
    • AdaBoost: weighted resampling and exponential loss
    • Gradient boosting: additive trees and descent along the loss gradient
  2. 60
    Key hyperparameters
    • n_estimators, learning_rate and max_depth
    • Balancing shrinkage and overfitting
  3. 61
    XGBoost
    • Tree regularization (L1/L2) and pruning
    • Column subsampling and tree learning
  4. 62
    LightGBM
    • Histogram-based binning and leaf-wise growth
    • GPU and multithreaded acceleration
  5. 63
    CatBoost
    • Ordered boosting and preventing target leakage
    • Native handling of categorical features
18 Anomaly detection 2 lessons
  1. 64
    Isolation Forest
    • The random partitioning principle
    • Path-length-based anomaly score
    • Key hyperparameters: n_estimators, max_samples
  2. 65
    One-class SVM
    • Kernel-based novelty detection
    • The ν parameter and the decision boundary
    • Sensitivity to scale
19 Boosting and anomaly detection lab 2 lessons
  1. 66
    Gradient boosting practice
    • Training and tuning XGBoost and LightGBM
    • Comparing learning curves
  2. 67
    Anomaly detection practice
    • Building Isolation Forest and One-Class SVM models
    • Evaluating with ROC and precision-recall metrics
20 Clustering: K-Means and hierarchical 3 lessons
  1. 68
    K-Means clustering
    • Centroid initialization and update steps
    • Choosing k with the elbow and silhouette methods
  2. 69
    Hierarchical clustering
    • Agglomerative and divisive approaches
    • Linkage criteria: single, complete, average
  3. 70
    Interpreting the dendrogram
    • Cutting the tree to define clusters
21 Dimensionality reduction: PCA 3 lessons
  1. 71
    Principal component analysis (PCA)
    • The covariance matrix and eigendecomposition
    • Explained variance ratio
  2. 72
    Dimensionality reduction workflow
    • Centering and scaling data
    • Choosing the number of components
  3. 73
    Applications and visualization
    • Preprocessing for ML and 2D/3D plots
22 Recommender systems 4 lessons
  1. 74
    Introduction
    • What it is and why it matters
    • Key use cases (e-commerce, streaming, news)
  2. 75
    Collaborative filtering
    • Memory-based: user–user and item–item similarity
    • Model-based: matrix factorization (SVD/ALS)
  3. 76
    Content-based filtering
    • Item profiles (metadata) and user profiles
    • Similarity via TF-IDF or embeddings
  4. 77
    Hybrid systems
    • Weighted and switching strategies
    • Example: metadata and viewing history at Netflix
23 Neural networks and deep learning (AI) 3 lessons
  1. 78
    What is artificial intelligence?
    • The differences between artificial intelligence, machine learning and deep learning
    • Modern AI systems and what they can do
  2. 79
    An overview of neural networks
    • Biological inspiration and historical development
    • Discussion: “Can machines really think?”
  3. 80
    Neural network fundamentals
    • The perceptron, activation functions and backpropagation
    • Building and training a neural network with PyTorch
    • Introduction to CNN and transformer architectures
24 Large language models (LLMs) and generative AI 3 lessons
  1. 81
    LLM fundamentals
    • Tokenization, embeddings and the transformer
    • Working with LLM APIs and prompt engineering
  2. 82
    RAG (Retrieval-Augmented Generation)
    • Document chunking and embeddings
    • Vector databases and semantic search
    • Building and evaluating a RAG pipeline
  3. 83
    LLM agents
    • Tool calling and agentic workflows
    • Security and guardrails in LLM applications
25 MLOps and model deployment 2 lessons
  1. 84
    Deploying the model
    • Saving the model and building a prediction API with FastAPI
    • Containerization with Docker
  2. 85
    Managing the ML lifecycle
    • Experiment tracking and the model registry in MLflow
    • Model monitoring and data drift
26 Project handover and interview 5 lessons
  1. 86
    Defining a clear business or research question
  2. 87
    Acquiring, cleaning and exploring data to understand patterns
  3. 88
    Preparing features and selecting suitable models from those covered (e.g. linear models, trees, boosting, SVM, etc.)
  4. 89
    Training, tuning and evaluating models using best practices (cross-validation, learning curves, the right metrics)
  5. 90
    Turning the workflow into a reproducible pipeline and presenting a short report summarizing the approach, results and lessons learned

Our instructors

Meet our experienced teachers!

  • Natiq Məmişov — instructor

    Natiq Məmişov

    Artificial Intelligence Engineering (AI Engineer)

    Don't know where to start?

    Not sure where to start?

    Get free advice

    Fill in the details and we'll get in touch with you.

    +994

    Your data is used only to contact you.

    Contact us directly (+994 10) 234 65 56
    CourseArtificial Intelligence Engineering (AI Engineer)
    • Address
    • 609 Jafar Jabbarli St., Baku / Globus Center

    • Social networks