Data Science Training with Real-Life Cases: San Francisco

course, attendees will proceed from theory to expert-led hands-on practice that encompasses a set of real cases to solve. In addition, you can submit a use case of choice to develop the expertise needed for your current business concerns. 

Who should attend?

Engineers who want to gain expertise in machine learning tools and frameworks
Everyone willing to move from theory to applied knowledge across challenging business tasks

Course objectives

Get the structured information you would otherwise have to look for in different sources
Explore the machine learning–related issues the practitioners face and the best practices to address them
Get ready-to-use scripts as the basis for creating algorithms of your own to solve business-specific problems
Collect valuable insights on the complete development life cycle of an ML solution
Get a fully-applicable template of the development life cycle, as well as recommendations for its subsequent adaptation to a changing business environment
Each trainee will have 16 hours of online Machine learning practice with a personal trainer on the project of your choice.

Program
DAY 1
Core Concepts and Techniques
Comprehensive review of the concepts, methods and models on which machine learning is based. In this module you'll learn:

Formal notation about ML tasks and definitions

Core principles of building an ML algorithms

Whole set ML algorithms, from Linear Regression to Random Forest

Introduction to core Python packages for ML

We'll cover the algorithms:

Linear and Logistic Regression

kNN and k-Means

Decision Trees and Random Forest

We'll show how to handle classification, regression and clustering tasks.
DAY 2
Feature Engineering and Development Methodology
Proven to work recipes and methods that help build better models and develop whole solution. We'll get a hold on a wide range of questions related to building ML models, such as:

Feature Engineering

Dealing with Missing Data and Outliers

Dealing with Imbalanced Classification

Advanced Validation Schemes

Handling of Versioning of models

CRISP-DM as main ML development methodology

DAY 3
Tabular Data
Transactional data and structured data sources in general are largely prevalent types of datasets, especially in telecom/banking. Purpose of this module is to show an approach for this data to retrieve useful insights.

Data preparation of transactional data

Time series specific family of algorithms

Statistical and Neural Network approaches for this task

PRACTICE
16 hours of hands-on practice

Real Estate Price Forecasting. Using the historical data of the Russian housing market along with demographic data, we will learn how to build a model for forecasting a house price.

Customer Income Prediction. We propose to analyze the customer data set in the Google Merchandise Store (also known as GStore, where Google Swag is sold). The goal is to create a model that predicts store revenue per customer.

Assessment of loan applications. This is a classic banking task to minimize financial risks. Using the client’s historical data, we will build a model that predicts the probability with which the client will return a bank loan.

Your own project. Each trainee can propose a project they'd like to work on.

At the end of the course, all participants receive a certificate of attendance. This certificate includes the training duration and contents, and proves the attendee’s knowledge of the emerging technology.

Prerequisites
Altoros recommends that all students have:
- Basic Python programming skills, a capability to work effectively with data structures
- Experience with the Jupyter Notebook applications
- Basic experience with Git
- A basic understanding of matrix vector operations and notation
- Basic knowledge of statistics
- Basic knowledge of command line operations
All code will be written in Python with the use of the following libraries:
- Pandas/NumPy are the libraries for matrix calculations and data frame operations. We strongly recommend to browse through the available tutorials for these packages, for instance, the official one.
- scikit-learn
- Matplotlib
All these libraries will be installed using Anaconda.
Requirements for the workstation:
- A web browser (Chrome/Firefox)
- Internet connection
- A firewall allowing outgoing connections on TCP ports 80 and 443
The following developer utilities should be installed:
- Anaconda
- Jupyter Notebook (will be installed using Anaconda)
If software requirements cannot be satisfied due to the security policy of your employer, please inform us about the situation to find an appropriate solution for this issue.

Meet the trainer
Kanstantsin Buzanouski, Data Scientist
Bio: Kanstantsin is a Data Scientist who strives to solve business problems with a defined plan at all stages of the development process. He has considerable hands-on experience in using Machine Learning and statistical methods in various domains as well as solving business problems starting from problem definition up to model fine-tuning and solution deployment. He is profoundly knowledgeable about current trends and approaches to Machine Learning. Kanstantsin works with popular frameworks and packages in both R (tidyverse + caret) and Python (pandas/numpy + scikit-learn). He constantly follows the latest trends in Machine Learning and never stops investigating new techniques and technologies to expand his skill set.
Vladimir Starostenkov, Machine Learning Architect
Bio: Vladimir has 10+ years of experience in software development. Over the course of his career, he has been part of 15 successful project implementations. Vladimir specializes in artificial intelligence and machine learning, distributed systems design, NoSQL and Hadoop-based systems benchmarking, permissioned blockchains,  data engineering, and development of data-centric apps. As an expert in NoSQL databases, he has authored a number of research papers, comparing the performance of Apache Cassandra, Redis, MongoDB, and Couchbase. Vladimir also serves as a trainer and a data evangelist. He is responsible for analyzing requirements, preparing training materials, and conducting training sessions. Vladimir is an active member of the Open Data Science Community.

Payment info:
If you would like to get an invoice for your company to pay for this training, please email to training@altoros.com and provide us with the following info:

Name of your Company/Division which you would like to be invoiced;

Name of the person the invoice should be addressed to;

Mailing address;

Purchase order # to put on the invoice (if required by your company).

Please note our classes are contingent upon having 5 attendees. If we don't have enough tickets sold, we will cancel the training and refund your money one week prior to the training.Thanks for the understanding.