From Neural Nets to Hashmaps — Why I’m Relearning the Fundamentals
Hashmaps power everything from feature stores to caching layers. In this post, I explain why I’m revisiting them as a machine learning engineer.
Stefan Nieuwenhuis
Welcome! I’m Stefan Nieuwenhuis, a Machine Learning Engineer building and documenting real-world systems to deepen my skills and share what I learn.
This blog is split into three main series:
- 💡 Recommender System – building a real-time recommender from scratch with Docker Compose, Spark, Redis/Faiss, and FastAPI.
- ⚙️ Homelab – setting up a bare-metal Kubernetes cluster with refurbished hardware.
- 🧠 DSA Refresh – revisiting data structures and algorithms with a focus on problem solving and pattern recognition.
Everything is built publicly, with full transparency into tradeoffs, failures, and wins. Follow along — or contribute!
Recent Posts
Recommender From Scratch: Building, Monitoring, and Learning in Public
In this series, I’m building a real-time recommender system from scratch, starting locally with Docker Compose.
Planning My Bare Metal K8s Homelab: Hardware, Goals & Learning Path
I’m building a Kubernetes homelab with:
Refreshing My DSA Knowledge: Why, How, and What’s Next
As part of my ongoing learning journey, I’ve decided to revisit fundamental data structures and algorithms.