The difference between data science and algorithm development

As a data scientist, you don't develop new methods, you apply them. Most data scientist use scikit learn, xgboost, keras, tensorflow, pytorch...

Algorithm development falls either in research or software development.

If it's in research then you need to come up with new ideas about algorithms. If it's in software development, you implement use cases of research ideas. Sometimes you can mix them both.

As a general rule, data scientists solve problems using pre-existing algorithms. They don't need to have experience in high-performing computing or software development. No need to write optimal algorithms.

A data scientist has many general skills, such as

• data collection
• software developer
• data engineering and cleansing
• algorithm development, selection and refining **
• statistical analysis
• data visualization
• etc.

A deep learning algorithm developer would be more specialized, concentrating more on that one aspect of the data science field, and less on the other skills.

Typically an algorithm developer will be part of a data science team, consisting of members having different skills, but overall the team will possess all of them.