Today I learned

  • Takaways From Making Sense Of Mvp

    Posted on January 20, 2026

    I appreciate blog post “Making sense of MVP” by Henrik Kniberg as it addresses a few important points when it comes to shipping customer solutions. In my professional carrer I have encountered already a few times issues along these points: [Read More]

  • Query Translation For Rag

    Posted on November 10, 2025

    Conceptually, RAG is simple: You’ve got your LLM but it lacks private information, corporate data or has a data cut-off which is too far in the past. Therefore, you make more data available, usually usig a vector store. [Read More]

  • Oauth 2

    Posted on October 16, 2025

    Up until today OAuth2 was always very opaque to me. Sure, I was using short-lived token to e.g. make API requests but I never tried to understand the mechanism itself. And actually it’s not that complicated from a conceptual point of view. [Read More]

  • Data Transformation Design Pattern

    Posted on May 15, 2025

    When transforming data, I have often been using a pattern like the following. I’ve created a class that is responsible for different data transformation steps. Generally speaking, structuring data transformations using classes makes sense as classes are a neat way to group related functions together. For another developer who has... [Read More]

  • Thrive Vs Protocol Buffers Vs Avro

    Posted on January 24, 2025

    When systems exchange data there are some obvious choices like csv, JSON or XML. The advantage of these are the human readability and wide adoption amongst different languages and frameworks. [Read More]

  • B Trees

    Posted on December 21, 2024

    B-trees are the most common way of index implementation in almost all relational database systems. [Read More]

  • Spark Distributed Write Pattern

    Posted on October 17, 2024

    When writing data with Spark, it’s easy to get confused as there are different ways to do so. The different ways do different things under the hood and depending on what you want it might be good or bad. This illustration helps to understand better what is going on when... [Read More]

  • Hiplot

    Posted on October 15, 2024

    Today I learned about a new python module called HiPlot. It uses a simple yet powerful visualization technique to uncover patterns and relationships in a data set. Uses cases are for example to identify the relevant features for a ML algorithm, to obtain decision rules for classification or to analyze... [Read More]