Sr. Data Scientist reveals: How to succeed in the Age of AI? How to become a data scientist? AI Products and more.
Lessons learned from Florian Roscheck (Senior Data Scientist @ Henkel)
AI is everywhere, right? New tools, new headlines every single day. It’s cool, but let's be real, it can also feel like a locked room only geniuses can enter.
I sat down with Florian Roscheck. Get this: Florian started out studying wind turbine engineering. Now? He’s a Senior Data Scientist over at Henkel, doing really interesting stuff in their R&D, and people are noticing – he was even named a top corporate influencer in Germany. Quite the path.
Our conversation wasn't what I expected – in a good way. We skipped the surface-level tech talk and got into what actually makes AI projects work or fail. If you build stuff, lead teams, or are just trying to figure out where this AI train is really heading, stick around.
Takeaways
Diverse backgrounds are essential in data and AI careers.
AI technology is becoming more accessible and standardized.
Data ownership is crucial for competitive advantage.
Implementation of ideas is more valuable than the ideas themselves.
Identifying the right problems to solve is key in AI.
User-centric design is critical for successful data solutions.
Networking plays a vital role in career advancement.
Transitioning to data science requires showcasing skills effectively.
Understanding AI is necessary for effective application.
Innovation in AI is driven by business needs. Authenticity is crucial for success in job interviews.
Understanding the business context is essential for data scientists.
Networking should be genuine and not just for job hunting.
Feedback, even when harsh, is an opportunity for growth.
Vulnerability fosters trust in professional relationships.
Future-proof roles in AI include product discovery and architecture.
Effective communication bridges the gap between technical and business teams.
Self-awareness through personality tests can guide career paths.
Embrace diverse experiences to create a unique professional niche.
Continuous learning and adaptation are key in a rapidly evolving field.
AI Tools Are Everywhere. So Now What?
Florian pointed out something huge right off the bat: the actual AI technology? It’s becoming easier for anyone to get their hands on. Think AutoML, easy-to-use APIs for the big language models. The tools are becoming standard issue.
So, if everyone can use AI, how does anyone actually get an edge?
Florian nailed it down to two things:
Your Own Data: Especially the stuff that’s unique to your world, your specific knowledge. Nobody else has that.
Picking the Right Problems: And this... this was the core of it.
The Big Question: Are We Just Building Tech Toys?
Florian’s been spending a lot of time thinking about product development and agile methods, and he sees a big disconnect. We can build almost anything with AI now. But who’s deciding what to build? Often, it’s the engineers – who are brilliant at building! – but maybe not always focused on finding the most valuable problem first.
He kept coming back to needing a product mindset, even if you're deep in the code. It’s less about “Can we build it?” and more about:
What problem are we really trying to fix here?
Is the flashy new AI thing (like a chatbot) actually the best way to fix it for the user?
Are we just shipping features ('outputs'), or are we changing something for the better ('outcomes')?
He mentioned using things like the Value Proposition Canvas – basically, mapping out what users need and struggle with before you even think about solutions. It sounds simple, but it's how you avoid building stuff that’s technically clever but doesn't actually help anyone or solve the real issue.
How a Wind Engineer Ended Up Here: Florian's Story
Florian’s own journey shows how mixing skills pays off. How’d he do it?
He Was Always Tinkering: Coding since he was a teen, messing with Excel VBA and microcontrollers.
Learning by Doing (and Helping): Took online courses, sure, but also helped a PhD student who needed coding help, built models for a startup.
Showing His Work: Put his side projects and learning on GitHub. That mattered.
Right Place, Right Time, Right Skills: He applied for a wind analyst job. The hiring manager saw his GitHub profile and all the extra stuff he’d been doing and said, "Hey, we're also looking for a data scientist, haven't announced it yet... interested?" Boom.
His advice now for people trying to get into data science felt really grounded:
Find Your Angle: Don't try to be good at everything. Specialize in a tool, a type of problem (like image recognition for chemistry), or being the bridge between two areas (like AI and finance).
Connect with People: Networking isn't just collecting contacts; it's finding people you click with. Good things happen from there.
Be You: Authenticity counts.
Sweat the Small Stuff (on your CV): If your resume is sloppy, it makes people wonder if your analysis will be too. Details matter.
Talking Tech to Non-Tech People (Without Losing Your Mind)
Florian splits his time 50/50 between tech and business folks. We talked about how hard that can be. His key?
Listen more than you talk.
Seriously. Understand what the business side is actually dealing with before you jump in with tech solutions. And yeah, sometimes you have to be the one to say, "Hold on, that AI idea sounds cool, but it might not be realistic or the best use of our time." It's part of the job to manage the hype.
Seriously, You Should Watch This
We talked about career shifts, why just being smart isn't enough, how projects really succeed, and getting practical about applying AI. Florian has this great way of mixing deep tech knowledge with real-world common sense.
If you're trying to make sense of AI, build something meaningful, or figure out your own path, you'll get a lot out of the full conversation.
Want the whole story? Grab a coffee and watch it here.
Let me know what you think!
Let’s Connect:
Mario Truss: LinkedIn | Website | YouTube
Today’s Guest Florian Roscheck: LinkedIn | Website | YouTube