Three Must Dos for a Successful Data Science Career | by Karun Thankachan | Aug, 2022
Opinion
Key things to do to fast-track your data science career
Data Science was declared the sexiest job of the 21st Century. With the excitement around AI, and publicity data science received, there has been an increasing number of folks entering the field every year. With such competition, it might seem difficult to stand out and achieve success in Data Science. This post covers some tried and tested actions to fast track your Data Science career, and no, it’s not another technical training or certification.
Data Scientists are tasked with creating value for a customer. The more value you can create, the more valuable you are to your company. There are three different levels of value you can create.
Low-level Value Creation: At this level, business stakeholders have stated an issue they want to solve. A senior data scientist on your team has identified how this issue can be resolved, identified the data required, and worked with the stakeholders to determine how the solution will be evaluated. The only thing left is implementing that solution. Sounds familiar? Well, this is what most entry-level data scientists work on. Building models/improving existing model where most of the problem-solving ambiguity has been removed. The only ambiguity left is the implementation ambiguity, which has minimal room for value-creation.
Mid-level Value Creation: At this level, business stakeholder have stated an issue they want to solve, and you as the Data Scientist work with them to resolve the issue. You identify what data is needed, what solution needs to be developed, and what metric makes sense from a business perspective to evaluate the model. Then you go about implementing a minimum viable solution and iteratively improving it to create value for business stakeholders.
High-Level Value Creation: At this level, due to your familiarity with the data and business domain, YOU APPROACH BUSINESS STAKEHOLDERS with a proposal to create value. At this level, you are essentially altering how business functions. If your proposal gets picked up, you follow through on developing it (as discussed earlier). This is the highest level of value creation and ideally the level you want to work at to fast-track your career.
Action Item — So how do you do this? Use the 80–20 rule. Spend 80% of your time working on what you are tasked with. Spend the remaining 20% understanding your business domain and the data you have access to. This means setting up one-on-ones with business stakeholders to understand what they work on, spending time understanding the available data, and reading up on trends in your business domain to spark ideas.
Document everything, and I do mean everything. Data Science is a multiphased process, and there is specific documentation that you need to maintain for each phase.
Proposal Phase: At this phase, a scientist works on clarifying the business problem, why it’s an important problem to solve, an estimate of the quantitative/qualitative benefit solving the problem could create. This is documented in a Research Proposal, created collaboratively with business stakeholder. The proposal could also contain the description of a minimum viable solution with a well-defined scope, how the solution will be evaluated, and an estimation of effort.
Development Phase: At this phase, a scientist works on identifying the data required to solve the problem and researching how a solution can be developed. As such, a scientist must work with data engineers to develop Data Documentation, describing what data was pulled and from where, how is data stored/secured, and how it will be updated. In addition, a scientist will develop a Literature Review, documenting existing approaches to solve the problem he is tasked with, identifying what a baseline solution is, and listing out promising approaches that can improve upon the baseline. Based on literature review, a scientist will experiment with different approaches and develop one or more minimum viable solutions to review with business.
Productization Phase: Once a minimum viable solution has been confirmed, a scientist works with an engineering team to decide on the architecture for training/deploying and serving the minimum viable solution. The entire process is documented in a Design Document.
Publicizing Phase: Once the solution is ready for release, it is good to support the solution with Publicly Accessible Documentation (like a Wiki Page). The documentation should list in layman terms what the problem is, why it’s important to solve it, what was developed, and what benefit it brought. If possible, create a video recording and upload it as well.
“[Your] important career decisions are made when you are NOT in the room” — Caroline Dowd-Higgins
The best way to improve your chances of making such decisions work for you is by having people who are familiar with your work be a part of that room. As such, you need to talk about your work.
Start by publicizing your work in newsletters/blogs/video channels within your company. Ask your manager to highlight it in team/leadership meetings.
Next, identify key product owners in your space who might be interested in your work. Set up a meeting to ask for feedback. This helps in two ways. One, the product owners get familiar with your work and will talk about it with others. Two, you get valuable feedback on your work and possible ways in which it could add additional value.
Finally, set up a meeting with seniors scientists on your team to ask for feedback. This again helps in the two ways discussed earlier. One, your senior scientists become aware of your work. Second, you get technical feedback on how you can improve the model, and your skill set!
To summarize, the three things you must do to fast-track your career are.
- Create high value for your company
- Document the work that you do
- Talk about your work and seek feedback
Sounds simple? The best things usually are!
Opinion
Key things to do to fast-track your data science career
Data Science was declared the sexiest job of the 21st Century. With the excitement around AI, and publicity data science received, there has been an increasing number of folks entering the field every year. With such competition, it might seem difficult to stand out and achieve success in Data Science. This post covers some tried and tested actions to fast track your Data Science career, and no, it’s not another technical training or certification.
Data Scientists are tasked with creating value for a customer. The more value you can create, the more valuable you are to your company. There are three different levels of value you can create.
Low-level Value Creation: At this level, business stakeholders have stated an issue they want to solve. A senior data scientist on your team has identified how this issue can be resolved, identified the data required, and worked with the stakeholders to determine how the solution will be evaluated. The only thing left is implementing that solution. Sounds familiar? Well, this is what most entry-level data scientists work on. Building models/improving existing model where most of the problem-solving ambiguity has been removed. The only ambiguity left is the implementation ambiguity, which has minimal room for value-creation.
Mid-level Value Creation: At this level, business stakeholder have stated an issue they want to solve, and you as the Data Scientist work with them to resolve the issue. You identify what data is needed, what solution needs to be developed, and what metric makes sense from a business perspective to evaluate the model. Then you go about implementing a minimum viable solution and iteratively improving it to create value for business stakeholders.
High-Level Value Creation: At this level, due to your familiarity with the data and business domain, YOU APPROACH BUSINESS STAKEHOLDERS with a proposal to create value. At this level, you are essentially altering how business functions. If your proposal gets picked up, you follow through on developing it (as discussed earlier). This is the highest level of value creation and ideally the level you want to work at to fast-track your career.
Action Item — So how do you do this? Use the 80–20 rule. Spend 80% of your time working on what you are tasked with. Spend the remaining 20% understanding your business domain and the data you have access to. This means setting up one-on-ones with business stakeholders to understand what they work on, spending time understanding the available data, and reading up on trends in your business domain to spark ideas.
Document everything, and I do mean everything. Data Science is a multiphased process, and there is specific documentation that you need to maintain for each phase.
Proposal Phase: At this phase, a scientist works on clarifying the business problem, why it’s an important problem to solve, an estimate of the quantitative/qualitative benefit solving the problem could create. This is documented in a Research Proposal, created collaboratively with business stakeholder. The proposal could also contain the description of a minimum viable solution with a well-defined scope, how the solution will be evaluated, and an estimation of effort.
Development Phase: At this phase, a scientist works on identifying the data required to solve the problem and researching how a solution can be developed. As such, a scientist must work with data engineers to develop Data Documentation, describing what data was pulled and from where, how is data stored/secured, and how it will be updated. In addition, a scientist will develop a Literature Review, documenting existing approaches to solve the problem he is tasked with, identifying what a baseline solution is, and listing out promising approaches that can improve upon the baseline. Based on literature review, a scientist will experiment with different approaches and develop one or more minimum viable solutions to review with business.
Productization Phase: Once a minimum viable solution has been confirmed, a scientist works with an engineering team to decide on the architecture for training/deploying and serving the minimum viable solution. The entire process is documented in a Design Document.
Publicizing Phase: Once the solution is ready for release, it is good to support the solution with Publicly Accessible Documentation (like a Wiki Page). The documentation should list in layman terms what the problem is, why it’s important to solve it, what was developed, and what benefit it brought. If possible, create a video recording and upload it as well.
“[Your] important career decisions are made when you are NOT in the room” — Caroline Dowd-Higgins
The best way to improve your chances of making such decisions work for you is by having people who are familiar with your work be a part of that room. As such, you need to talk about your work.
Start by publicizing your work in newsletters/blogs/video channels within your company. Ask your manager to highlight it in team/leadership meetings.
Next, identify key product owners in your space who might be interested in your work. Set up a meeting to ask for feedback. This helps in two ways. One, the product owners get familiar with your work and will talk about it with others. Two, you get valuable feedback on your work and possible ways in which it could add additional value.
Finally, set up a meeting with seniors scientists on your team to ask for feedback. This again helps in the two ways discussed earlier. One, your senior scientists become aware of your work. Second, you get technical feedback on how you can improve the model, and your skill set!
To summarize, the three things you must do to fast-track your career are.
- Create high value for your company
- Document the work that you do
- Talk about your work and seek feedback
Sounds simple? The best things usually are!
Denial of responsibility! Techno Blender is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – [email protected]. The content will be deleted within 24 hours.