Architectures for ordered data — text, time series, audio — culminating in the transformer, the architecture behind modern LLMs. Part of the ML Engineer track.

  1. RNNs, LSTMs & GRUs — [Advanced] — recurrent networks that process sequences step by step, and how gating fixed their memory problem
  2. Transformers & Attention — [Advanced] — self-attention, multi-head attention, and why transformers replaced RNNs
  3. NLP & Embeddings — [Advanced] — tokenization, stemming/lemmatization, and word/sentence embeddings
  • Deep Learning — the fundamentals these build on
  • How LLMs Work — the applied view of the transformers this section explains from the modeling side
  • RAG & Embeddings — embeddings put to work in applications
↑ Back to 02 — ML Engineer