← Tech Station
01 — Start here

AI Primer

The concepts every LLM practitioner needs before writing a single prompt.

19 guides in this track·1.6h reading·Foundations
What you'll learn

Build the mental model — how models learn, generate text, and where their capabilities end — so every downstream decision (prompting, RAG, fine-tuning) rests on real understanding, not folklore.

Who this is for

Engineers, PMs, and analysts who use LLMs daily but have never opened the box. No ML background required — working knowledge of software is enough.

How to use it

The guides are numbered — read in order for the curriculum path, or jump straight to the one you need. Each card is self-contained.

The guides

Ordered by recommended reading path
01
Fundamentals of AI, ML, Deep Learning & Gen AI

The nested hierarchy of the field and how each term relates.

5 minRead →
02
How Machines Learn: Data, Training & Generalisation

The shift from rules to learning-from-examples; training vs inference.

6 minRead →
03
What Is NLP? Teaching Machines to Work with Language

The language problem, why it is hard, classic vs modern approaches.

5 minRead →
04
Representing Text as Numbers

Why models need numeric input; one-hot, bag-of-words and their limits.

4 minRead →
05
Tokenization: How Text Becomes Tokens

Tokens vs words, subword tokenization, why tokens are the unit of cost/context.

5 minRead →
06
Word Embeddings & Vector Space

Meaning as geometry; similar words as nearby vectors.

5 minRead →
07
Neural Networks: Neurons, Layers & Weights

The basic building block, forward pass, how layers stack.

5 minRead →
08
Activation Functions & Softmax

Non-linearity, common activations, and how softmax turns scores into probabilities.

5 minRead →
09
How Networks Learn: Loss, Gradient Descent & Backpropagation

The training loop that adjusts weights to reduce error.

6 minRead →
10
Sequence Models & Why RNNs Struggled

Handling ordered text and the bottlenecks that motivated attention.

5 minRead →
11
The Attention Mechanism

Letting a model weigh which other tokens matter.

5 minRead →
12
Self-Attention & Multi-Head Attention

How tokens attend to each other; why multiple heads help.

5 minRead →
13
The Transformer Architecture, End to End

Assembling embeddings, attention and feed-forward layers into the full model.

5 minRead →
14
From Transformer to LLM: Pretraining at Scale

Next-token prediction over massive data; emergence of capability.

5 minRead →
15
Fine-Tuning & Alignment (SFT + RLHF)

Turning a raw predictor into a helpful assistant.

5 minRead →
16
How LLMs Generate Text: Sampling & Temperature

Decoding, probability distributions, the creativity/reliability dial.

5 minRead →
17
Context Windows & Tokens in Practice

What the context window is, why it is limited, how it shapes design.

5 minRead →
18
Why LLMs Hallucinate

The plausibility-vs-truth root cause and what it implies.

6 minRead →
19
Capabilities, Limits & Where This Is Heading

Honest map of what LLMs do well, do poorly, and what is next.

6 minRead →

Related tracks

Continue in the same group
02 — Core skillPrompt EngineeringZero-shot to few-shot, chain-of-thought, and structured output patterns.03 — ArchitectureRAGChunking, embeddings, vector search, and fixing hallucinations in production.04 — LLMOpsLLM Fine-Tuning & LLMOpsWhen to fine-tune vs. prompt, LoRA, QLoRA, evaluation, and deployment.05 — AdvancedAgentic AIPlanning loops, tool use, memory, and multi-agent coordination.