Notes from talking with data scientists
Last July was posted an article from Harvard Business Review on data scientists with few advices on the role’s evolution (https://hbr.org/2022/07/is-data-scientist-still-the-sexiest-job-of-the-21st-century): “We expect to see continued differentiation of responsibilities and roles that all once fell under the data scientist category. Companies will need detailed skill classification and certification processes for these diverse jobs, and must ensure that all of the needed roles are present on large-scale data science projects. Professional data scientists themselves will focus on algorithmic innovation, but will also need to be responsible for ensuring that amateurs don’t get in over their heads. Most importantly, data scientists must contribute towards appropriate collection of data, responsible analysis, fully-deployed models, and successful business outcomes”. Those guidelines confirm my current objectives which are 1) know how to build a scaled product, 2) know algorithms that business don’t and 3) focus on solving a business issues. Nevertheless it made me think on what others are doing to stay up-to-date with data science and how they prepare their future. Therefore, I interviewed five data scientists from work and highlighted what I thought were the most interesting responses.
Keeping up to date with machine learning
- Check the top ML conferences from few months ago, and look at the key papers which generated a lot of talks. You wait few months to make sure the paper is still relevant and is worth spending some time on. You don’t want to dive into the math as it takes too much time (except if it is very important), and focus instead on the core ideas. A good tip is to search for the keynotes speeches, which tend to give a good overview without diving into technical details
- Follow key AI players on SNS and see what they do and talk about. Using LinkedIn and Twitter were suggested
- Keep working on core data science skills, whether it is statistics, probability, math or algorithms!
- Leverage the possibilities offered at work. Many companies will sponsor employees to attend (virtually or not) key conferences. Another way is to organize regular meetings where data scientists / researchers where you can get exposure to what others are doing. For example, each week a member can present a new ML paper; ideally, the paper should be generic enough that most people can find it interesting or relevant to their work. This is a good way to get a general idea of what people are doing in other domains such as NLP, vision or speech processing
- Write down who are the key data scientists in your department and organize from time to time some 1 to 1 meetings where you casually discuss what they do, which algorithms they work on, etc. I can confirm that some 20 minutes “catch up” meetings often got me way more interesting information than monitoring data science news online
- To keep up with algorithms, several data scientists mentioned using platforms like Youtube to understand key concepts rather than reading the original paper. Besides, it is more enjoyable and faster
Networking
- Not many people I talked to were taking specific actions toward networking. Someone mentioned using LinkedIn and Medium to reach out to new people, asking them questions, etc. Another way is to join your city ML group (meetup or else)
Mentoring
- None of the data scientists I talked to had mentor but some definitely wished they had. Apparently, getting mentor in data science is not an easy task. To overcome this problem, they have experts in their field to go-to when needed (someone for statistics, another one for ML questions, etc…).
Staying up to date technically
- Writing blog article is nice option to try new techniques and save what you learned
- Reading books with practical examples on how to solve data science / ML problems
- All data science jobs don’t require to constantly chase better technical skills. Prototyping / research jobs are a good example of that; however, you need to train from time to time for code interviews to stay up-to-date
Future jobs or plan
- To my surprise, the majority of data scientists I talked to were interested in future management positions. One motive is higher salaries and another reason is to have more impact to solve business problems. To follow up with the latter, one data scientist was considering starting a MBA in order to be “part of the business”, rather than staying limited to something too technical. I think it highlights that to be as impactful as possible you need to speak the business language and understand their problematic. In addition, I think data scientists tend to often focus at first on their programming and statistics / ML skill while improving business knowledge comes after
Strategy to stay ahead of curve
- Stay up to date with code interview: some recommendations are the book ‘Cracking the code interview’ and a lot of Leetcode exercises https://leetcode.com/
- Strengthen your fundamentals (statistics, math, programming…) and try to work in other fields, such as A/B testing, NLP, data engineering, etc.
- Do Kaggle competitions
