Today marks the end of my chapter as a software engineer at Rockefeller, where I spent the past year working on enterprise machine learning systems. Before this, I studied Data Science and spent time as a developer at startups in Paris and San Francisco. I found myself moving between software engineering, data science, and product. It was not indecision as much as trying to understand where the real leverage was. What I learned is that I care about being close to decision-making, where engineering isnāt abstracted from impact.
Rockefeller forced a kind of clarity I didnāt have before. I began to see the difference between ādata scienceā as a major and machine learning as a software engineering discipline.
That reflection led to the lessons below.
1. The Blurriest Job of the 21st century
In 2012, Harvard Business Review called Data Science āthe sexiest job of the 21st century.ā
Since then, the role has stretched in every direction:
- PhDs developing research-grade models
- Business analysts doing dashboards and SQL
- Startup hires asked to āfind insightsā for high-growth products
The truth? What data science means entirely depends on the environment.
At school, datasets are clean and questions are well-defined. In industry, neither the data nor the problem is. You might find yourself in a position where you have to prove your own value to your team. The real metric is adoption, not accuracy.
2. Data Science ā Data Engineering ā ML Engineering
What most people call āmachine learning workā is actually two different worlds:
Experimentation | Engineering |
Modeling in a notebook | Making the model real, reliable, and used |
What is taught in school | What creates value in industry |
And the engineering side is where most of the work and impact lives. It is where you can be in the top 1%.
3. The Path That Made Sense for Me
Step 1 : Data Engineering (Foundation)
This was my bread and butter, and where trust is built.
- Reliable ingestion pipelines across heterogeneous sources
- Schema enforcement & validation rules
- Lineage tracking & reproducibility
- Understanding database architecture
- Cloud architecture (Azure / AWS / GCP)
This is where I was able to provide value right away. If the data isnāt stable, nothing above it can stand.
Step 2 : ML Engineering (Scale & Durability)
This is where models become systems.
- Deployment workflows & CI/CD
- Model versioning and lifecycle management
- Latency & throughput tuning
- API design and integration
- Monitoring, drift detection, failure mode handling
This is where top engineers differentiate themselves. It requires real software-engineering depth and is harder to break into than data engineering, which made it a natural next step for me.
Ā
4. Mantras
- Write code that is easy to delete.
- A model that canāt survive changing conditions is not a production modelā¦ itās a demo
- Models should generalize, not just perform well on historical data
- Overfitting vs. generalization are operational challenges not just theory
- Subject matter expertise is not optional. Interpretability gets you a seat at the table with executives.
- Build a system that works when you are not in it
- Technical work teaches you patterns; people teach you principles.
My role at Rockefeller was about alignment. Teams, trust, and clarity move the code forward just as much as design patterns do. Thatās what the past year made clear to me, which takes me to the next point. The most important thing.
5. It is the people, stupid
It is about who do you know and who knows you. Early in your career, the single best thing you can do is find excellence fast. Good does not even know what great looks like. Find great people, and they will show you. At Rockefeller it was clear thinking, ownership, speed, humility.
It really is the people, but also the words that attract them. Narrative is your shortest path to bending reality. In early careers, people often wait to āearnā clarity before talking about their story. Iāve found it works the other way around: you speak your direction into existence. When you talk openly about what youāre building and what youāre curious about, the right people find you.
Interestingly, the offer I received for my next role was the result of the CEO messaging me on LinkedIn, saying heād enjoyed one of the blogs I wrote on my personal site. Obviously, it wasnāt the blog that landed me the job. It was the culmination of effort over time - projects, writing, and curiosity made public. The blog was just a signal and my website is a mirror of what Iād already been building quietly for years.
As I move into my next chapter, Iām excited to stay close to both engineering and business impact while working with the best. More on that soonā¦ Onwards!
Ā
Resources I recommend:
Navigating different career paths in Tech? roadmap.sh
Interested in learning about careers in ML and AI? https://www.youtube.com/@MarinaWyssAI