One year from doubting that AI assistants were any use to me and worrying about what they cost in energy, to using one every working day, told in order from my own record, with the doubts that stayed. [read]
2026-10-11 / AI / ai, machine-learning, learning, craft
A short narrative review that frames machine learning as data, representations and models, lists opportunities and drawbacks, and ends with three recommendations. Useful as a pointer to its sources; weak as a source itself. [read]
2026-10-11 / reading, machine-learning, chemical-engineering
A thirty-three page review of AI in process systems engineering with 174 references, a catalogue of reported results by function, a barrier list, and a roll call of industrial deployments. Every number in it is secondhand and the industrial evidence rests on company pages. [read]
2026-10-11 / reading, machine-learning, process-systems-engineering
A history of machine learning in chemical engineering along three threads, models, data and applications, with a four-step workflow, three application classes and an outlook that argues general materials databases cannot carry process problems. [read]
2026-10-11 / reading, machine-learning, chemical-engineering
A short opinion piece on end-to-end molecular models, graph networks, transformers and matrix completion, for pure and mixture properties, with practical guidance on when they beat UNIFAC and COSMO-RS and a route for carrying them into process simulators. [read]
2026-10-11 / reading, machine-learning, thermodynamics, process-design
A fitted model can forecast an output or be asked which inputs it relies on, and this is the vocabulary for the second use with the limits its own documentation gives. [read]
2026-08-23 / AI / series From zero / machine-learning, interpretation, feature-importance
A SHAP value splits one prediction of one fitted model among its inputs, so it describes the model, and there are four common readings of the plot that it does not support. [read]
2026-02-10 / AI / series From zero / shap, interpretability, machine-learning
What a kernel is, what C, epsilon, gamma and the length scale mean in support vector regression and Gaussian process regression, why the inputs are scaled first, and what the uncertainty from a Gaussian process rests on. [read]
2025-11-25 / AI / series From zero / machine-learning, support-vector-regression, gaussian-processes, scikit-learn
Twelve words from the scikit-learn workflow, each explained by a run table or a calibration curve, with the name the library gives it. [read]
2025-11-25 / AI / series From zero / machine-learning, scikit-learn, vocabulary
What a decision tree, a random forest, gradient boosting and XGBoost are, what their main settings mean, and why none of them can predict beyond the range of the runs they were trained on. [read]
2025-11-25 / AI / series From zero / machine-learning, decision-trees, xgboost, scikit-learn
What the units, layers, activation and epochs of a multilayer perceptron are, why its inputs are scaled first, and why the fit on a small table changes from one start to the next. [read]
2025-11-19 / AI / series From zero / neural-networks, scikit-learn, machine-learning
The first days of learning Python and machine learning as a chemical engineer with no coding, with the questions the assistant asked, the plan it gave and what two evenings covered. [read]
2025-11-19 / AI / python, machine-learning, learning-to-code