THE PATH
The full journey, end to end. Get a personalized roadmap →
Zero Knowledge
Programming, Linux, math fundamentals.
Foundations
Python, NumPy, linear algebra, calculus.
Machine Learning
Optimization, regularization, evaluation.
Deep Learning
PyTorch, tensors, neural networks, training loops.
Transformers
Attention, RoPE, decoder-only architecture.
Tokenizers
BPE, SentencePiece, vocabulary design.
Tiny LM
Train a character-level model from scratch.
Small LLM
10M to 100M parameter transformer.
Fine-Tuning
SFT, LoRA, QLoRA, DPO.
Evaluation
Benchmarks, perplexity, contamination.
Optimization
Mixed precision, memory, throughput.
Multi-GPU
Data parallelism, FSDP, ZeRO.
Deployment
Inference, serving, quantization.
Source Library
Papers, docs, and specs behind everything above.