There's a lot written on how to break into data science. There's a lot written on new techniques. There isn't a lot written on how to become a better data scientist. Being a data scientist is so much more than machine learning algorithms and techniques. Techniques are tools we use to reach a larger goal: positively impacting the business. Learning the technical skills is just the first step.

From reading Sequoia Capital's levels of a data scientist and from conversations with my manager, I've started to paint a picture of what the progression of a data scientist looks like. By painting this picture, this gives me a path to progress.

Across the industry, the levels of an IC data scientist generally look like this:

Level 1 - Execution: A level one data scientist typically will typically:

Technical - Capable of applying different ML algorithms to different problems, manipulating data in Python, SQL and running basic statistical analysis with some oversight.

Independence - Primarily focused on execution in tandem with a Senior DS or Manager.

Problem Formulation - Will require a well-defined problem from a business partner or manager.

Influence/Leadership - Percieved to have high standards of execution, with error-free analysis. Business partners trust this person to produce quality work.

Level 2 - Independence:

Technical - Capable of producing an end-to-end ML solution with little oversight from Sr DS, while also contributing new technical tooling for the team.

Independence - Can independently lead communication with business partners and own this area with little assistance.

Problem Formulation - Capable of taking a loosely defined problem from the business and break it down into an execute-able roadmap.

Influence/Leadership - Capable of influencing decision-making with knowledge of the product/part of the business.

Level 3 - Mentorship:

Technical - Applies more advanced techniques/concepts to multiple problems while writing elegant, re-producable code. Should be capable of writing their own packages.

Independence - Tend to be paired with a junior DS to be a mentor. Strong knowledge of multiple areas, which allows them to capture synergies between the business.

Problem Formulation - Develops the project roadmap to hit goals set by executive teams.

Influence/Leadership - Thought partner with Product Managers on area of expertise. Work side-by-side with business partners to develop roadmap.

We assume that each subsequent level will contain all the skills of the previous level.

For me to progress, it's important to do a gap analysis on my current skills. By identifying the gaps, I can develop a plan that specifically fills these gaps.

Technical Skills

Areas for Improvement:

• Producing error-free analysis, significantly reducing the frequency of bugs (Level 1).
• Understand how to increase the speed of my code so I can provide results quickly(Level 1).

Influence/Leadership

Areas for Improvement:

• Become a domain expert within a specific product area (Level 2).
• Begin influencing decisions within a product (Level 2).

Problem Formulation:

Areas for Improvement:

• Take a vague, undefined problem and break it down into a concrete roadmap (Level 2).

Independence:

Areas for Improvement:

• Independently handle conversations with business partners.

Conducting my own gap analysis, the most important thing for me is to reduce the frequency of bugs within my code and increase the speed at which I produce results. By reducing the frequency of bugs and producing results faster, this will build trust with business partners. With higher trust, this will open up opportunities to take on more vaguely defined projects. In addition, having more influence comes from having more knowledge on a particular domain, which means I'll need to pick a side of the business to hone in on.