Every idea, told three ways.

Machine Learning Fit a function to data: supervised when the labels exist, unsupervised when you are hunting for structure without them, self-supervised when the data…

Neural Networks Layers of weighted sums pushed through non-linearities.

Gradient Descent Differentiate the loss with respect to every parameter, then step against the gradient.

Loss & Error MSE for regression, cross-entropy for classification, and a long tail of task-specific objectives, but the design decision that matters is that the…

Features & Encoding One-hot, ordinal, target and hashing encoders.

Evaluation & Metrics Hold-out sets and cross-validation for honesty.

Python Slow in the loop, unbeatable as the glue.

NumPy & pandas Vectorised arrays and dataframes: the discipline is thinking in whole-array operations instead of row loops, knowing where copies happen versus…

scikit-learn One consistent interface – fit, predict, transform – over most of classical ML, plus pipelines that keep preprocessing honest between training and…

PyTorch & TensorFlow Tensors, autograd and the training loop: define the forward pass, and the framework records the computation graph and differentiates it for you.

Notebooks & Tracking Jupyter for exploration, MLflow or Weights & Biases for runs, parameters, metrics and artefacts.

Data Pipelines Batch and streaming, orchestrators and retries, schema drift, late-arriving data, backfills that rewrite history, and idempotency so a rerun does not…

Large Language Models Transformers and attention, pretrained on next-token prediction at a scale that turns a party trick into reasoning-shaped behaviour, then aligned…

Tokens & Embeddings Subword tokenisation sets your context budget and your bill – token counts, not word counts, are the unit of everything.

Prompting & Context Instructions, examples, retrieved documents, tool results and conversation history all compete for one finite context window, and the model attends…

Retrieval (RAG) Chunk the documents, embed them, retrieve by hybrid search (dense vectors plus keyword – dense alone misses part numbers and names), rerank, then…

Fine-tuning & Adapters LoRA and the PEFT family: small trainable deltas over frozen weights, so you tune millions of parameters instead of billions and swap adapters per…

Generative AI at Work The generation call is the easy tenth of the system.

AI Agents A model plus a loop plus tools plus state, and the model is the smallest part.

Loops & Graphs A loop decides at runtime – model in charge, maximum flexibility, unbounded variance.

Tools & MCP Function calling with schemas, and MCP emerging as the standard for exposing tools across vendors – write once, plug into any model.

Memory Three different problems wearing one name: working context (this session), episodic notes (what happened before), durable facts (what stays true).

Reinforcement Learning Agent, environment, reward, policy.

RLHF & Preference Train a reward model on pairwise human preferences, then optimise the policy against it – PPO classically, the DPO family when you would rather skip…

Evals & Hallucination Task-level suites, graders (exact-match where you can, LLM-as-judge where you must, with the judge itself evaluated), golden sets guarded from…

Guardrails & Responsible AI Least privilege for something that improvises: scoped credentials, allow-lists over deny-lists, approval gates on irreversible actions, sandboxes for…

MLOps & Deployment Versioning for models, data and prompts together – a prompt change is a deploy.

Customer Communications Template and content management, variable data, composition engines, channel orchestration across print, email, SMS and app, plus archive and…

Content & Documents Capture, classification, extraction and lifecycle over enterprise content, the classic ECM stack, with document intelligence turning scans into…

Cost, Latency & ROI Route easy traffic to small models and hard cases up the tier.