We invite you to training from developers of Intel Software and Yandex "Software Development in the Cloud"

On October 11 and November 7, Intel Software Solutions holds a free master class for developers in the Moscow office of Yandex on Leo Tolstoy. Specialists of both companies will tell everyone about software tools and cloud technologies for software development, high-performance computing and machine learning.

Consider that we have already invited you. Do not forget to register only - the number of seats is not infinite.



We invite everyone:

• software developers, researchers, scientists or engineers;
• working on projects for processing a large amount of data or developing software in the cloud;
• knowing C, C ++ or Python.

At the training, you can listen to the following master classes:

Cloud CI and Intel Software Tools in the Cloud

Intel Software Tools is a must-have tool for a software developer who seeks to maximize the performance of his code on Intel Architecture (x86) platforms. Intel software tools provide compilers, runtimes and libraries, as well as profiling and prototyping tools covering the full range of optimizations - from algorithmic to microarchitectural.



As part of the master class, we:

• try compilers, runtimes and libraries;
• discuss CI setup for using Intel tools;
• consider the problems and their solutions.

Cloud Application Performance Analysis

Migrating applications to the cloud infrastructure offers several advantages - for example, easy scaling, redundancy, technical support, and lower TCO. At the same time, working with cloud services requires a competent and attentive team of software development and maintenance.

The purpose of the workshop was to uncover the capabilities of the Yandex.Cloud infrastructure (based on Intel Cascade Lake) and use the Intel Parallel Studio XE Professional Edition performance optimization and prototyping tools.

Introducing the CatBoost Library

CatBoost is an open source gradient boosting library.

It surpasses analogues in quality and has additional advantages: for example, it supports working with categorical attributes (music genres, device IDs, URLs) without data preprocessing.



We will get acquainted with the capabilities of the library on the example of the classification problem. We will go through all the stages of building a forecasting model and consider the following topics:

• selection of suitable loss functions and metrics for optimization;
• model training;
• visualization of the learning process and cross-validation;
• work with the built-in retraining detector;
• selection of the optimal decision threshold;
• the importance of features and interpretation of model predictions;
• applying the trained model to test data.

You will find all information about the training on the invitation page .



The event starts at 10:00. Participation is free upon prior registration . Remember to bring your laptop with you!