Machine Learning and Kotlin: A Practical Guide

The book "Machine Learning and Kotlin" is the first practical guide in Russian that shows how to build machine learning models using the Kotlin and JVM ecosystem, without needing to write code in Python. The author, Alexey Voronov, Senior ML Engineer and Kotlin enthusiast, shares years of experience integrating ML into production Kotlin applications.

What is this book about?

The book covers the complete cycle of working with data and models in Kotlin:

• Machine Learning Basics — linear regression, decision trees, SVM, clustering. Theory and immediately code in Kotlin.
• ML Libraries for JVM — detailed analysis of KotlinDL (an analog of Keras for Kotlin), Smile (statistics and ML), DeepLearning4J (neural networks). Performance comparison.
• Working with Data — loading, cleaning, visualization using Kotlin libraries (kravis, lets-plot). No need for Pandas.
• Deep Learning — convolutional and recurrent neural networks, transformers. Examples: image classification, text analysis.
• ML in Production — packaging models into REST API (Ktor, Spring Boot), exporting to ONNX, working with MLflow.
• Kotlin Multiplatform for ML — running models on Android, iOS, and the server.

Who is this book for?

• Java/Kotlin developers who want to integrate machine learning into their projects but don't want to learn Python and its frameworks.
• Data Scientists looking to master Kotlin as an alternative for production code.
• Students and teachers studying ML on JVM.

What will you learn after reading?

• How to choose the right library for a task: from simple regression to GPT-like models.
• How to read and process data without Python (CSV, JSON, Parquet, databases).
• How to build a neural network in KotlinDL in 10 lines of code.
• How to optimize a model for JVM and use GPU via CUDA.
• How to deploy a model in a microservice on Ktor and connect it to the frontend.

Why Kotlin for ML?

Kotlin is a statically typed language running on JVM. It combines Python's conciseness with Java's performance. Using Kotlin for ML, you get:

• A unified technology stack (backend + ML + mobile applications).
• Fewer type errors at compile time.
• High execution speed (JIT compilation).
• Easy integration with existing infrastructure (Spring, Ktor, Gradle).

The book contains over 50 practical examples, ready to run. The source code is available on GitHub. Suitable for both beginners and experienced ML engineers.

Format: PDF, EPUB. Pages: 480.