Self-Study Plan For Moving From A Junior Data Scientist To A Senior Data Scientist
You are familiar with the basics of data science and now you want to level up.
You're familiar with applying off-the-shelf ML algorithms and have gotten your feet wet with data wrangling and messy datasets. Now you want to go beyond where you are now and improve your data science skills. Unfortunately most guides, FAQ, and articles you've encountered are ways to dive into data science not on how to go beyond the basics.
To go beyond the basics, you need to look at what it takes to be hired as a senior data scientist.
You want to level up from being a beginner data scientist. The most efficient way to figure how to level up is to look what it would take to get hired at job levels higher than where you are. To do this look at job boards and figure out what you are missing. Anything you are missing should be put into your self-study plan.
Let's look at an example to drive this point home.
Let's say that you are already a junior data scientist want to level up to a senior data scientist position in the next few years. Great, go to a job board (indeed.com is used in this example, you can also use a country specific job board) and type in "senior data scientist". Here are some results -> https://www.indeed.com/jobs?q=senior+data+scientist&l= ....
The top result for when we just ran this search was this job (saved as a PDF in case it disappears). The PDF is found here -> Example Senior Data Scientist Job Posting PDF. This positions requests 6-9+ years of experience, so this should be a good example to learn from.
From this result, you can see that as a "Senior Data Scientist", on any given day you would be doing some of the following:
Using your deep knowledge of numerical and statistical packages (Pandas, Numpy, Sklearn, R) to...
- Implement a gradient-boosting classifier to predict whether a person is likely to visit a car dealership based on the advertising signals they’ve received.
- Use a Bayesian dynamic time series model to estimate the causal impact of an advertising campaign on sales at your neighborhood grocery.
- Model the complex interactions between system architecture components to refactor and rethink key components and models in an advertising system.
- Develop algorithms to optimize the setting of every lever in our advertising infrastructure.
- Analyze data to better understand how a neighborhood’s consumption of web pages correlates with visits to a local big box store.
- Build a time series model to forecast future sales of diapers for one of our clients.
- Model the effects of environmental changes on promotion effectiveness with multiple regression.
Writing complex database queries using distributed computing frameworks: MapReduce, Hadoop, Impala and Spark to establish links between large datasets in order to...
- Find the handful of outliers in billions of transactions.
- Evaluate competing bidding models for RTB auctions to inform our bid pricing strategy.
- Feed data into your latest ensemble model aimed at maximizing the return on a client's online advertising budget.
Leveraging your experience with real world data to...
- Derive a set of new features that will help us better understand the interplay between geography and audience features to improve our model performance.
- Discover and explore third party data sources to determine their value for improving our model performance.
- Build new data-driven products and bring them to market.
Provide technical leadership to...
- Mentor other data scientists in algorithms, models, tools, and products that make the team more efficient.
- Participate in planning, roadmap, and architecture discussions to help evolve our data science into revenue-generating products.
- Engage in code and model reviews to continually raise the bar on our work.
- Draw data flows and architecture designs on the white board to encourage understanding and cohesive development towards your solution.
- Meet with customers and help map business needs into product requirements.
Great - you now have a list of advanced stuff you need to be able to do in order to get hired as a senior data scientist at this company. Looking at the list, most of these things would also get you a senior position at most other companies. So how do you learn this stuff? Google (or your favorite search engine).
Looking at the very first bullet point "Implement a gradient-boosting classifier" all you need to do is to add the words "how to" in front of the key technical term and search for it on Google. The first result for me was this article -> http://tullo.ch/articles/gradient-boosted-decision-trees-primer/. If you can read the article, know every word, be able to derive all the terms, and code it up in your favorite programming language, then you will have gone beyond the "basics" and be on your way.
You want to level up from being a beginner data scientist in the most effective and efficient way possible.
Unfortunately most guides, FAQ, and articles you've encountered are ways to dive into data science not on how to go beyond the basics.
Take a deep breath and remember that the most efficient way to figure how to level up is to look what it would take to get hired at job levels higher than where you are. To do this look at job boards and figure out what you are missing. Anything you are missing should be put into your self-study plan.
Your next action to level up from a junior data scientist to a senior data scientist...
Your next action is to first look at three senior data scientist job postings you are interested in and find the knowledge gaps you have from where you are now to where you'll need to be. This way you'll learn something new that will help you get one step closer to getting leveling up to a senior data science role.