Blog

Notes on applied machine learning.

Practical writing on production ML, explainable AI, RAG systems, model evaluation, and the engineering details that decide whether a model survives contact with real data.

September 28, 2026

Building and Evaluating a RAG System in Five Steps: An EU AI Act Example

A step-by-step walkthrough of a question-answering system over the EU AI Act: scoping it, writing the test first, measuring retrieval, checking every answer, and changing one thing at a time, with the results of each step.

RAG RAG Evaluation LLM Evaluation Retrieval-Augmented Generation EU AI Act Information Retrieval Reranking Golden Dataset Python

September 28, 2026

Building and Testing an AI Research Agent in Five Steps: An EU AI Act Example

A step-by-step walkthrough of a research agent that writes cited reports on the EU AI Act: choosing how complex it needs to be, giving it the right sources, scoring claims in code, adding a reviewer, and testing that reviewer with planted errors.

AI Agents LangGraph LLM Evaluation Multi-Agent Systems EU AI Act Claude Python

September 15, 2026

Explainable AI for Graph Neural Networks: Example Use of CAM and Grad-CAM

Class activation maps are almost always explained on images. This post applies CAM and Grad-CAM to a spatio-temporal graph network for human skeleton data, with the math, PyTorch code, and the architecture detail that decides whether the two methods give the same answer.

Explainable AI Graph Neural Networks Graph Convolutional Networks Class Activation Map Human Pose Estimation Grad-CAM PyTorch Skeleton Data Human Activity Recognition Human Action Recognition Deep Learning

July 4, 2026

Signal Quality Assessment in Python: Are Your PPG, ECG, EMG, or EEG Signals Clean Enough to Use?

Before calculating the heart rate or HRV from your time series data, you need to know whether the signal is ready for further processing or modeling. This post shows a one-metric quality check for four common physiological signals, tested on real data.

Signal Processing Biomarkers Python Wearables PPG signal quality ECG signal quality EMG signal quality EEG signal quality signal quality index

June 6, 2026

Can Fake Training Data Make AI Harder to Fool? A Traffic Sign Experiment

A traffic sign experiment on whether synthetic training data can improve adversarial robustness without sacrificing accuracy on real images

Computer Vision Synthetic Data Robustness AI safety Generative AI