Nikolay Eremeev
Senior Data Infrastructure Engineer. I build the pipelines and platforms that carry data at scale — Flink, Spark, Kafka, Kubernetes — and the LLM automation that now runs on top of them.
- German/EU citizen · based in Germany, CET
- Full-time, contract or advisory
- Email is the channel I actually watch
About
Full CV →Senior Data Infrastructure Engineer at CrowdStrike, previously IBM and TeamViewer. Real-time pipelines, cloud cost, CI/CD and IaC across AWS and GCP, and LLM-powered automation running in production. Published research on context selection for LLM code review (Zenodo).
Platforms Flink, Kafka, Spark, Kubernetes, Terraform
Clouds AWS, GCP
Location Germany
Writing
All articles →- Software Engineering Principles: from SOLID to the AI Era
A reference of software engineering principles with examples in Java, Rust, Go and Python 3.12+ — from SOLID and parse-don't-validate to distributed data, streaming and systems with an LLM inside.
- Spark DAG Scheduler: From Logical Plan to Physical Execution
How Spark's DAGScheduler turns an action into jobs, stages and tasks: where shuffles cut stages, ShuffleMapStage vs ResultStage, retries, a clickable DAG.
- CSR: How Graphs Are Stored in Flat Arrays
Interactive visualization of Compressed Sparse Row: three flat arrays replace nested hash maps for sparse graphs, for 10-30x speedups from cache locality.
- What is a matroid? Definition, examples, and why greedy works
What a matroid is, in plain terms: the exchange axiom, graphic, linear and uniform examples you can click, and why greedy is optimal exactly on matroids.