My Data Science Internship Experience
Coming into it, I wasn’t sure what to expect from a remote internship experience—in light of the pandemic, I was grateful to still have an internship with GSI Technology at all. Now, a quick three months later, it’s almost over! So to wrap-up, I thought I’d share some of my takeaways from my data science internship at GSI.
I knew I was going to be writing blogs for GSI from seeing the past interns’ work but I wasn’t sure how to even begin that process. Luckily, I began with a pretty focused topic for my first blog and my mentor helped me make an outline to guide my research. A writing process that ended up working for me was to do a bit of preliminary, general research on the topic of interest such that I could jot down an outline, and then to focus my research further based on the outline; otherwise, the sheer amount of articles and papers covering these data science concepts could quickly get overwhelming.
Speaking of research papers, I got a lot of experience reading them this summer! Specifically, I gained confidence in knowing that if I spend enough time grappling with papers I’ll generally be able to make sense of them, something that’s important to keep in mind when faced with all of the technical jargon and complex equations in data science literature—it can be pretty intimidating initially.
I quickly came to terms with the fact that there was no way for me to fully grasp any of these data science topics within a matter of weeks, as they are subjects people spend their entire careers researching, and I learned to be okay with developing a decent surface level understanding of things, so long as I understood enough to write effectively about them. After all, I wasn’t writing to experts in the field necessarily but to people who were at similar experience levels to myself prior to when I began researching, which brings me to perhaps my most important takeaway from blog writing: know your intended audience. Knowing my audience and keeping in mind the central messages I wanted to convey to them helped me keep my research focused and avoid going down unnecessary rabbit holes (not to say that didn’t still happen, occasionally 🙂).
I also learned to refer to outside resources for direction when I couldn’t find the answers I was looking for online: the authors of papers or GitHub repositories, industry experts, etc. Most of the time, I found people are very willing to provide advice and suggestions — the worst thing that can happen is they don’t reply, and the best thing is that they share information that saves you a lot of time and/or strife in the long run.
In sum, I learned a lot about current state-of-the-art research in various realms of data science over the course of my internship at GSI. I did a deep dive into satellite image processing and image segmentation, scratched the surface of reinforcement learning, and researched techniques for model compression and task retention in neural networks. And blogging about these topics really helped solidify my knowledge of them and gave me practice communicating technical concepts in an easily interpretable way to people who may be less familiar with them—a skill that will definitely be of use in future workplaces. Finally, I’m really grateful to my mentor, George (gwilliams@gsitechnology.com), for his expertise, patience, and guidance along the way.
With that, the following is a list of the blogs I’ve published during my internship at GSI, marking the end of a great summer and a valuable learning experience!
- Getting Started with Satellite Data Processing
- A Beginner’s Guide to Segmentation in Satellite Images
- Demystifying Markov Decision Processes
- Core Concepts in Reinforcement Learning By Example
- An Overview of Model Compression Techniques for Deep Learning in Space
- Your Neural Network Will Forget What It’s Learned!
