Hi, I'm Dennis. I'm a software engineer and here I write about things I dig into: Mostly AI/ML, Wasm, and Java. I share what I learned to help others and my future self. You can read more about me on the About page.
Fixing HTTP 502 errors of Services Behind an Application Load Balancer
Aligning the timeouts between hops can prevent random 502 Bad Gateway errors.
Hitting the ceiling of an ML model
Noisy data cannot be balanced out by changing parameters of a model.
Selected reading
- AI/ML Implementing Retrieval-augmented generation (RAG) with an own LLM
- AI/ML Client-side inference using onnxruntime
- Wasm Under the surface of WASI in Rust
- Wasm Creating an own WASI function
- Software Engineering Hexagonal Architecture
- Software Engineering Imperative vs. declarative programming
- Java How Spring MVC handles requests
- Java Java CRaC
- Java Preventing TOCTOU attacks
MCP might be a better interface for internal tools
MCP with its available clients might serve better client interfaces for internal tools.
Random links to understand Machine Learning and AI
Some links I found nice and would like to share
Usage of Git worktrees
A short intro to Git worktrees and how they help manage parallel tasks in the same repository without constantly switching branches.
Client-side inference using onnxruntime
ONNX Runtime Web runs ONNX models in the browser for client‑side inference.
5 practical use-cases for an ML-powered RSS digest
Beside creating an ML model for an RSS digest the other fun way is to actually use it. Here I take a look at the different possible use-cases.
Optimizing a model by using metrics
This post shows how to use metrics for optimizing a model and how it helped to improve the AUC from 0.59 to 0.84.