AdvancedProduction ML architecture: training-serving skew
The model is the small part. Training-serving skew, feature freshness, shadow deploys, drift detection, and monitoring that catches decay.
Tag
Machine Learning
Every article tagged Machine Learning. For a structured path, browse Learn.
AllAiAnnouncementsArchitectureCodingComputer VisionDatabasesDeep LearningErrorsHand TrackingLearningLlmMachine LearningMediapipeNumpyOpencvPandasProductPythonRag
Advanced
AdvancedData leakage: the kinds that survive your train/test split
Splitting correctly stops the obvious leak. Target encoding, feature stores, label windows and proxy features leak anyway — how to find each.
AdvancedML model evaluation mistakes that make good numbers meaningless
Nine errors behind great offline metrics and bad production models: random splits on time series, threshold-free metrics, tuning on the test set.
AdvancedCNNs from scratch in NumPy: convolution as a matrix multiply
Convolution forward and backward with no framework. The im2col trick that makes it fast, why the backward pass is a convolution too, and a net that trains.
AdvancedTransformer from scratch in PyTorch, block by block
A working decoder-only transformer in 120 lines. Pre-norm, residual streams, why the MLP is 4x wide, weight tying, stable initialisation.
AdvancedThe attention mechanism, worked out by hand
Queries, keys and values with real numbers. Why the scaling factor is sqrt(d_k), what multi-head attention buys, and how causal masking works.
AdvancedEmbeddings mathematically: what cosine similarity measures
The geometry of embedding space: why you normalise, what dimensionality really buys you, and the anisotropy problem nobody mentions.
AdvancedLLM evaluation: grading a model with no right answer
Vibes do not scale and BLEU measures nothing you care about. Eval sets, LLM-as-judge without fooling yourself, and the biases that corrupt it.
AdvancedBuilding RAG properly: chunking, evaluation, reranking
Most RAG fails at retrieval, not generation. Build a golden set, measure recall@k, run chunking experiments, add a reranker when data says so.
AdvancedVector databases under the hood: HNSW and IVF-PQ
ANN search trades accuracy for speed. How HNSW graphs and IVF-PQ quantisation work, which knobs move recall, and how to choose an index.
AdvancedFine-tuning vs RAG: how to actually choose
Fine-tuning teaches behaviour, RAG supplies facts. A decision framework, the LoRA maths, real costs, and the cheaper ladder to climb first.
BeginnerWhat matrices actually do in machine learning
Your dataset is a matrix, a model is a matrix, and training is matrix multiplication repeated. Six operations, drawn out, with the ML job each one does.
IntermediateMatrix multiplication in NumPy: @ vs *, and the shape error everyone hits
Why A @ B and A * B give completely different answers, what "shapes not aligned" really means, and the one rule that makes matrix shapes click for good.
BeginnerJupyter vs VS Code vs Colab: where should you learn machine learning?
Free GPUs, offline work, real debugging — each tool wins somewhere. An honest comparison plus the setup that gets you the benefits of all three.
IntermediateClean a messy CSV with pandas before you train anything
Ten pandas fixes for the problems every real dataset has: broken headers, duplicate rows, mixed types, silly outliers and inconsistent categories.
BeginnerHow to detect overfitting in your own model
Overfitting means your model memorised the training data instead of learning from it. Here are the three checks that reveal it and the fixes that work, in order.
BeginnerConfusion matrix, precision and recall — without the maths headache
Accuracy hides the mistakes that matter. Learn to read a confusion matrix, tell precision from recall, and pick the metric that fits what your model is actually for.
BeginnerTrain/test split explained — and why your accuracy is a lie
A 99% accurate model that fails in production usually has data leakage. Learn the four ways test data sneaks into training and how a scikit-learn Pipeline prevents all of them.
BeginnerTrain your first machine learning model in 20 lines of Python
No maths degree required. Load data, split it, fit a model, check the score — the four steps every scikit-learn project uses, explained line by line.
BeginnerValueError: could not convert string to float — fix your data, not your code
This error means one cell in your data is not a number. Learn how to find the exact bad row in pandas and clean commas, currency symbols, percent signs and blanks for good.
BeginnerStop memorizing Python syntax
You don’t need every Python trick. Master these seven patterns and you’ll read — and write — almost any beginner ML notebook with confidence.
BeginnerWhat is Sythra?
Sythra is a machine learning education platform — structured courses, an AI tutor that teaches, in-browser labs, and Project Studio so you learn ML by building.