Machine Learning Engineer

What the job involves

A machine learning engineer turns machine learning models into working services. They prepare data, train and validate models, tune parameters, then package the model into an application or API that a product can use, and monitor its quality after launch. Such models power recommendation systems, image and speech recognition, text processing and forecasting. What sets the role apart from a data scientist is a stronger emphasis on software engineering and deployment.

Skills you need

• confident Python and data libraries
• maths: linear algebra, statistics, basics of optimization
• classic machine learning and the basics of neural networks
• frameworks such as PyTorch or TensorFlow
• SQL, Git, containers and the basics of deploying services
• the ability to evaluate a model and find the causes of its errors
• discipline and readiness to keep learning new things

How to start and learn

Start with Python, statistics and data analysis basics, then study classic machine learning methods and only afterwards neural networks. Consolidate each topic with a project on open data and publish it on GitHub. The course assumes a basic knowledge of programming and maths, so a beginner should first learn the fundamentals. Direct entry is less common than coming through a developer, data analyst or data scientist position.

Where they work

In Armenia machine learning specialists work in startups and product companies, outsourcing studios and international teams, banks and telecommunications, and as freelancers. There are relatively few positions, and employers usually expect a strong technical base.