The many aspects of Data Science
Andrey Lukyanenko
Senior DS, MTS AI
DS - many meanings
Courses vs. Real World. General challenges
- Sometimes the problem isn't well defined
- Not all projects need ML
- People don't always need your solutions
- Explaining models
- Convincing people
- Domain knowledge is good, but data is better
- ML often is a small part of a project
- Incorrect usage of predictions
Courses vs Real World. Not all data is good
- No data
- Not enough data
- Garbage in garbage out
- Leaks
- Different test data
- Bias or not representative
- Bad infrastructure
Courses vs Real World. More business problems
- Model predictions influence future models
- Coronavirus
- Unique projects
- Talent deficit for companies
Be aware of fallacies
Plan your career
- Balance between short-term and long-term
- Choose what you want to do
- Draw boundaries - responsibilities, overtime
- Life-work balance. Consider self-study time
- If you don't like dealing with the business side - work on products or try deep learning
- Learn how to say "no"
Develop your career
- Learn how to search for information
- Learn how to study efficiently
- Competitions and hackathons
- Pet-projects
- Hone basic skills. Especially soft skills
- Open-source
- Join communities, but not all of them
- Develop personal brand
Links
Contacts
Data Science aspects
By Andrey Lukyanenko
