https://ift.tt/jJ8hOxP Opinion A proposal of where to take your skills when you’ve reached the intermediate/advanced stage End of the Be...
Opinion
A proposal of where to take your skills when you’ve reached the intermediate/advanced stage
Part of the reason I’m writing this is to document my adventure and recent challenges. I’m a Senior Data Scientist and I have worked on a full range of DS/ML solutions from Causal Inference to traditional ML to Deep Learning. I’ve been so obsessed and heads down in my craft that I spent a large part of my journey so far mastering Data Science itself (stats, linear algebra, Python, etc.).
The application of these tools was always a bit secondary to me since I was laser-focused on becoming truly an expert in the tools themselves. After a lot of long days and countless hours, I find myself performing my craft with a high degree of confidence, just as I aspired. I’m looking for the next mountain to climb and it largely feels like sitting with all the tools in my hands searching for a problem to fix.
This story will explore this feeling, identify what I plan to do, and offer to grow together.
The Feeling of the Aimless Master Builder
There are many ways people can find Data Science and ML today. I get really fascinated by the people who specialize in fields completely out in left-field like Civil Engineering and then transition into Data Science to amplify their work. Personally, I started from next to nothing and only with Data Science.
I didn’t really come from a domain that I had extensive expertise in ready to apply Data Science methods. I had interests, but nothing I was intimately tied to. I found a lot of wonder in the field of Data Science and Machine Learning so I spent a lot of my time so far wide-eyed learning the intricacies of AI, Causal Inference, MLOps, and more. After finishing my Masters in Applied Data Science and working as a Senior Data Scientist for some time now, I’m left wondering what my “next mountain” looks like. Should I dive deeper into stats and become a statistical master? Should I become more of a software engineer and learn how to program at the highest ability? Or maybe shift focus to the product space to build data products? There are a number of mountains ahead that I am evaluating and they all sound similarly enticing.
This feeling feels a lot like knowing my tools extremely well and being highly skilled in using any number of them at a given moment, but searching for the right problem to solve. It can feel a little uneasy if I’m being honest. My instinct is to take online courses or degrees to give myself the feeling of accomplishment, but I don’t believe that’s really going to solve what I’m searching for which is purpose. For me, I’ve had to conclude that I’m looking to find the right niche domain to apply data skills in to solve challenging global problems. Finding that niche domain that I can call my own has been tough though.
Most of my career has been spent in this uncertainty so I feel right at home trying to put this puzzle together. If you’re reading this feeling similarly stuck, know you’re not alone.
The Corporate Routes
If you bring this thought process to your boss, you commonly are suggested two routes: the technical individual-contributor and the business managerial routes.
The individual-contributor route will ask you to dive far deeper into the excellence of the tools. Becoming a master of Python or statistical rigor or analytical speed is what is truly required down this path depending on your specialty. Cassie Kozyrkov has a phenomenal write-up discussing this a bit more from the analyst’s perspective. This will lead you to a Team Lead or a Principal type of role.
The managerial route will ask you to dive far deeper into the business of the company and/or the managing people and teams. Becoming a master of communication, data product management, and leading teams are typically what’s needed down this route. There are countless books and courses on how to be an effective leader out there, but as a data leader you definitely stand out as an outlier in today’s world (especially if you aspire to be a good one). It’s not an easy job to translate business needs and desires into a successful data project while also managing teams and having oversight over the technical components. There really aren’t many [useful] blueprints out there that show how to be an effective data leader today, which adds to the difficulty of this route.
There is potential here for true unicorns to be a little of both as well. It’s hard but if you have a high EQ and people skills and then choose to specialize deeply in your skillset then it can naturally fit you right in between the two spaces. The challenge with this hybrid is making sure you still have some balance and aren’t burning out.
There’s nothing wrong with any of the routes proposed above and if you’re already at the Senior/Lead level you’ll likely follow one of them regardless. With that being said, I want to propose two more routes that I think often get overlooked but are terribly crucial for expert data scientists to follow.
The Passion Routes
The majority of the [best] Data Scientists I know have come into this career with a deep passion for the craft. That passion fueled a lot of long hours and late-night studying to master their craft. When we hit the relative apex of our efforts in learning the rigor, the reality is that the best use of time for most isn’t reading another research paper to keep on the cutting-edge of the craft. We feel that mastering more math formulas or algorithms is going to make the same difference as it did at the beginning of our career, but in reality, we’re playing a different game now.
As beginners, it’s imperative to master foundations enough to excel in your craft. As experts, you want to excel in the application of your craft and this requires specializing in a domain of application. It’s more useful for the majority of the intermediate/expert data scientists to become masters in domains they deeply care about (or can evolve to care about). Are you an artist of some sort? Are you interested in trying to help fight climate change? Are you looking to go deeper or broader in the world of finance? It is important for you to specialize, but choosing to specialize in a domain of application is going to yield far greater rewards than learning the newest algorithm that has just been released. As a caveat, this doesn’t apply as well for the PhD candidates I feel. They can be expert data scientists and in many cases, they need to be domain and tool experts. Their application scope doesn’t usually affect business needs though and it remains in the research perspective.
Other than mastering a specific application avenue, content creation is becoming more and more of a realistic evolution for intermediate and expert data scientists. A digital creator or teacher is a great way to keep your skills sharp while also building a community around you. With sites like Medium, YouTube, and Twitter taking off for data creators you can get started with little overhead. This isn’t as easy to truly make it far in since there are a lot of people attempting it. Doing this successfully can require persistence and quality which demands a large volume of your time and effort. If you enjoy expanding your skillset and connecting with others, then I think it can be a really valuable experience to at least try out.
As with the corporate routes, I believe you can do both of these passion routes as well. They can take up a large amount of your time, but any of these routes would easily do that already. These aren’t meant to be “easier” than the climb-up most experts have already done so far. In fact, contrary to the corporate routes, I think doing both will actually make you more successful in each route for the passionate ones. If you choose a niche application domain that you want to commit to and also start building data content around it, you’ll become a more effective practitioner because you’ll attract like-individuals in that community while also diving deep really quickly.
My Recommendation: Choose the Passionate Routes
The reason expert Data Scientists are experts is not due to a high mathematical skill — it’s due to their grit. This grit comes from a passion that needs to be properly taken care of or it can easily whittle away. The corporate routes are going to be what most are forced to choose between already, and the passionate routes are more options than necessities. Even then, I highly recommend that you choose one or both of those options for yourself. I guarantee it is the best way to keep your passion burning brighter than ever and, more importantly, it’s what the community needs. We don’t need another “Become a Data Scientist in 6 Months!” course — we need real humans sharing their weaknesses and putting their grit and passion on public display. This showcasing of problem-solving for areas you care about is how we should be welcoming people into this field — learn the programming and math to solve problems you care about and the world needs to be solved.
In my own journey, I recently made a similar call. I started writing less than a year ago and have been searching for the application specialty that I want to dive far deeper into. It took a while to narrow it down but ultimately the environmental and sustainability space kept calling to me. I don’t have prior training in this domain so it’s been a lot of self-teaching, reading from experts, and working on a self-directed project to go further. I’ll start documenting and writing about my journey as this keeps evolving but this has been the most excited I’ve been about my career since I started first learning about AI. If there are other climate scientists who are looking to collaborate, share knowledge, or work on a project together please reach out! Would love to hear from you.
For others in a similar spot as me, this story from Maria Leis was exactly what I needed to jump right into the climate change arena. Stay tuned for future stories of my climate science projects and musings!
What’s Next for Expert Data Scientists? was originally published in Towards Data Science on Medium, where people are continuing the conversation by highlighting and responding to this story.
from Towards Data Science - Medium https://ift.tt/Ncq04V8
via RiYo Analytics
No comments