In the Author Spotlight series, TDS editors talk to members of our community about their career path in Data Science and AI, their writing, and their sources of inspiration. Today, we are happy to share our conversations with them Maria Mouschoutzi.
Maria is a data analyst and project manager with a strong background in application research, mechanical engineering, and maritime logistics optimization. He combines industry-in-industry experience translated into decision-driven analytics to develop decision-support tools, end-to-end processes, and communicate Insights across technical and non-technical teams.
between “What 'thinking' and 'reasoning' really means in AI and LLMS,” He talks about the semantic gap between human and machine thinking. How has understanding this distinction influenced the way you approach model development and interpretation in your professional work?
AI has generated a lot of hype recently. In no time, many ML-based products have been quickly repackaged as AI, and there seems to be a renewed demand for anything AI-focused on. Because of this, I believe that it is now important for everyone to have a basic understanding of the technology of what AI is and what it already exists, so that they are in a position to test it.
The truth is that we carry a lot of responsibility about the nature of AI, which comes from the narrative from our Sci-Fi. This property makes it easy to get carried away by all the exciting and promising possibilities and forget the real power that exists, in the end it is good as some kind of magic potion that will reduce all magical problems. Non-technical business users tend to overdo this with AI, sometimes seeing it as a dark superintelligence, capable of providing the right answers and solutions to anything.
For better or worse, this cannot be better in reality. LLMS – The main scientific achievement is what all the AI Fruss is about – very good at some things (for example, generating emails or summaries), but not good at other things (for example, analyzing multilevel relationships and multilevel relationships).
Having a technical understanding of what AI is and how it basically works has helped me a lot in my professional career. Basically, it allows me to find valid AI use cases and manage business user expectations of what can and cannot be done. At a more technical level, it allows me to isolate specific elements that need to be used in specific situations, so that the solution delivered has real business value.
For example, if a Rag application is required to search for certain technical documents and perform calculations based on the information found in that document, then the last part of the code needs to be included in the application to perform the calculations (instead of enabling the model directly).
Where do you draw the initial inspiration for your articles, especially more philosophies such as the “cool water” series of lectures?
The first thread of my “cool water cooler” talk comes from actual conversations I've witnessed in the office, as well as friends' conversations. I think that because of people's tendency to avoid unnecessary conflict in Corporate Setups, sometimes some ideas that are not really true can be expressed in normal discussions around the water cooler. And usually, no one calls out wrong facts just to avoid conflict or to challenge co-workers.
Even though such discussions are kind and well-intentioned – in fact they are just from work – sometimes they lead to the promotion of incorrect scientific facts. Especially in complex and not-so-good-sounding topics like math and AI, we can easily miss things and develop wrong ideas.
The first idea that inspired me to write a whole piece about it 'If you play enough rounds of roulette, you will eventually win, because the odds are about 50/50, and the results will begin to balance out.' Now, if you've ever taken a math class, you know this isn't how it works; But if you didn't have a category, and no one calls this, you can leave this discussion with some unusual ideas about how gambling works. So, my initial inspiration for that series was very abstract math topics.
However, the same – if not more – ambivalence applies from these topics to AI-related topics. The huge hype that AI has produced has resulted in people thinking and spreading all kinds of misinformation about how AI works and what it does with incredible confidence. That's why it's so important to teach them the basics, whether it's math, AI, or any subject.
Can you walk us through your typical writing process for a detailed technical article, from initial research to final completion? How do you balance deep technical accuracy with general audience availability?
Every technical post starts with a technical concept that I want to write about – for example, showing how to use a certain library or how to program a certain problem in Python. For example, in my Pokémon post, the goal was to explain how to program a research problem in Python. After identifying the basic technical concept that I want to focus on, my next step is usually to search for relevant data that can be used to illustrate.
I believe this is the challenging and time-consuming part – finding good, open source data that can be freely used for analysis. While there are many datasets out there, it doesn't hurt to find one that is freely available, complete with data, and interesting enough to tell a good story.
In my opinion, the flavor of the data you will use can have a big impact on the popularity of your post. Planning an Operations research problem using Pokémon sounds a lot more fun than using employee shifts (EWW!). Overall, the Dataset should match the technical topic I have chosen and make for a specific related topic.
Once you have identified the technical topic of the post and the data I will use, I write the actual code. This is a straightforward step: Write code using data and get it to work and produce the right results.
After I finish the code and make sure it works properly, I start writing the actual post. Usually I start my posts with a short intro to what started with my enthusiasm for this particular topic (for example, I wanted to make a complex life for my PhD, and that this Searoute Python course becomes useful) how to help you write calls to any API).
I also include general definitions, wherever appropriate, the basic premise of the use case I demonstrate, and a brief introduction to the code libraries I will be using.
In the main part of the technical post, I show how I edit the code with Python Snippets, and step-by-step explanations of how they play and the expected results that everything works well.
I also like to add GIF screenshots to show any active graphics included in the code – I believe they make the post more interesting, easy to understand.
And there you have it! Technical lesson!
What initially motivated you to start sharing your knowledge and understanding with the wider Data Science community, and does the writing process lead back to your professional work?
Back in 2017, when I was writing my diploma thesis, I stumbled upon the medium and data science for the first time. After reading more posts, I remember being completely overwhelmed by the amount of technical fabric, the variety of topics, and the violence of the posts. It felt like a data science community, with writers from different backgrounds and at different technical levels – there were articles for all levels and different backgrounds.
But apart from appreciating the technology of the tutorials that allowed me to learn and understand more about data science, I loved the creativity and storytelling in the post. Unlike the GitHub page or the overwhelming response, there was some creativity and creativity in many of the posts. I really enjoy reading such posts – they help me learn a lot about science and machine learning, and over time, I've quietly developed a desire to write such posts as well.
After thinking about it for a while, I came back and sent my first post, and this is how I publish TDs for the first time shly at the beginning of Tds, since then, I have received several posts from that first post.
One thing I really enjoy writing technical pieces for TDS is sharing things that I find challenging to understand or very interesting. Sometimes complex topics such as careers, opportunities, or AI can feel intimidating and intimidating, preventing people from starting to read and learn more about them – I am guilty of this.
By creating a simple, straightforward, seemingly sweet version of a complex topic, I feel like I'm allowing people to start reading and learning more about it with a gentle, not-so-good start and see for themselves that it's not so scary after all.
On the flip side, writing has helped me a lot on a personal and professional level. My written communication has improved a lot. Over time, it has become easier for me to present complex, technical topics in a way that a non-technical business audience can understand. Finally, you put yourself in the position of explaining the topic to someone else in simple terms that forces you to question it completely and avoid leaving meaningless spots.
Looking back on your career, what is a non-technical skill you wish you had focused on earlier?
In the data profession, the most important technical skill is communication.
While communication is important in any field, it is especially important in data roles. It's essentially what bridges the gap between hard technical work and practical business understanding, and helps make you a well-rounded data professional.
This is because, it doesn't matter how strong your technical skills are, if you can't communicate the value of your deliverables to business users and management, they won't take you very far.
It is important to be able to explain the value of your work to a non-technical audience, speak their language, understand what is important to them, and communicate the findings in a way that shows how your work benefits them.
Data and analytics, as important as they are, often feel too intimidating or confusing to business users. Being able to translate data into meaningful business direction and then leverage those insights effectively ultimately allows your data analytics projects to have a real impact on the company.
To learn more about Maria's work and stay up to date with the latest articles, you can follow her on TDS or on LinkedIn.
