L
learnllm
Back to Catalog
Free Track4.5 Hours•Level: Beginner•Updated: October 2026

LLM Foundations: Inference, Tokenomics & Architecture

Deconstruct transformers from the bottom up. Master Byte-Pair Encoding, KV caching, attention compute patterns, and deploy lightning-fast local inference clusters.

Start Track

Curriculum Outcomes & Engineering Deliverables

Calculate memory footprint for KV Cache and weights across FP16, INT8, and FP4
Trace autoregressive next-token prediction and speculative decoding loops
Inspect subword tokenization anomalies and special token hijacking attacks
Spin up production vLLM and Ollama instances with continuous batching

Included Engineering Deliverables

Custom BPE Tokenizer Trainer & Sanitizer Suite
Interactive KV Cache VRAM Mathematical Calculator
Production Docker Compose deployment for vLLM & Ollama
End-to-end Token Injection Safeguard Middleware

Curriculum Modules

3 Modules • 5 Lessons
Module 012 Lessons

Tokenization, BPE & Context Windows

How raw human text transforms into mathematical high-dimensional vectors.

Byte-Pair Encoding & Special Token Injection
Interactive • 24 min
Free
Multi-Head Attention & Rotary Position Embeddings (RoPE)
Deep Dive • 32 min
Free
Module 022 Lessons

The KV Cache & Autoregressive Decoding

Understand why LLM inference is memory-bandwidth bound rather than compute bound.

PagedAttention & Memory Footprint Calculations
Architecture • 28 min
Free
Speculative Decoding: 2.5x Speedup Without Quality Loss
Interactive • 32 min
Free
Module 031 Lessons

Hands-on Local Deployment

Step-by-step setup of enterprise vLLM clusters with continuous batching.

Deploying High-Speed Local Inference with vLLM & Ollama
Code Lab • 42 min
Free

Hardware & GPU Compute

Baseline GPUAny modern CPU or Consumer GPU (RTX 3060+)
VRAM Requirement6GB - 12GB VRAM
Recommended ConfigurationApple Silicon (M1/M2/M3) or NVIDIA RTX 4070 (12GB)

Prerequisites & Tech Stack

Stack Components
PyTorch 2.5vLLM 0.6TikTokenOllamaHuggingFace
Knowledge Prerequisites
  • Python fundamentals
  • Basic linear algebra (matrices & vectors)
  • Familiarity with CLI