read · learn · build
113 articles. 15 stages. One coherent path.
The theory side of the site — the full organized/
learning path, vendored for offline-friendly publishing. Each
article cross-links to the hands-on companion track that turns
the theory into running code. Browse below, or
open the curriculum view if you'd rather
have a starting line picked for you.
Editor's picks
Read me first
Top-level docs that frame the path itself.
Browse by stage
Math Foundations
Linear algebra, probability, calculus, info theory.
Ml Fundamentals
Supervised, unsupervised, evaluation, regularization.
Neural Networks
Backprop, activations, optimizers, regularization.
Language Modeling
n-grams to RNNs to why transformers won.
Tokens Embeddings
How text becomes vectors.
Transformers
Self-attention, multi-head, KV caching, GPT from scratch.
Modern Llms
Scaling laws, MoE, reasoning models, long context.
Prompting
Few-shot, CoT, structured outputs, sampling.
Rag
Chunking, embeddings, hybrid search, reranking, evals.
Fine Tuning
SFT, LoRA, RLHF, DPO, GRPO, embedding fine-tuning.
Data & Tooling
02Distillation
Embedding Fine-Tuning
04Field report: Llama 3 — frontier post-training, in 92 pages
05LoRA & QLoRA
Field report: Phi-3 — synthetic data and distillation, in the open
07RLHF, DPO, GRPO — Preference and Reward Training
Supervised Fine-Tuning (SFT)
09When to Fine-Tune
Agents
Loops, tools, memory, planning, multi-agent, browser/vision.
Multimodal
CLIP, VLMs, diffusion, video, speech, synthetic data.
Production
Deployment, evals, guardrails, observability, cost, data systems.
Applications
Text-to-SQL, code, browser agents, finance, case studies.
Engineering Career
Roles, learning roadmap, staying current.