Session 1
Session 1: Layers of Intelligence
AI, ML, Deep Learning & Generative AI fundamental layers
Session 2
Session 2: Deep Learning
Neural networks, training dynamics, loss functions & backpropagation
Session 3
Session 3: Language & Machines
NLP foundations, tokenization, stemming, POS tagging & NER
Session 4
Session 4: Embeddings & Language Models
Word2Vec, semantic vector math, TF-IDF & autoregressive models
Session 5
Session 5: Memory in Machines (RNN vs LSTM)
Sequential memory, vanishing gradients & LSTM gating mechanisms
Session 6
Session 6: The Evolution (LSTM to Transformer)
Self-attention mechanism, scaled dot-product & transformer revolution
Session 7
Session 7: Emergence of Generative AI, BERT & GPT
Probabilistic models, VAEs, GANs, Attention math, BERT bidirectional & GPT autoregressive generation
Session 8
Session 8: What Are Foundation Models & How Do Models Learn?
Architecture scale, emergent abilities, quantization/pruning efficiency, and pretraining paradigms (Supervised, Unsupervised & Self-Supervised)
Session 9
Session 9: Pretraining Paradigms, Fine-Tuning & RLHF Alignment
CLM vs MLM vs Contrastive learning, PEFT LoRA math ($W_0 + B \cdot A$), QLoRA 4-bit, SFT, and RLHF alignment via PPO & KL divergence
Session 10
Session 10: LLM Scaling (MoE, Reasoning, Context) & Evaluation Metrics
Mixture of Experts (MoE) router gating, DeepSeek-R1 Chain-of-Thought reasoning, million-token context extensions, and Intrinsic (Perplexity, BLEU, FID) & Extrinsic evaluation metrics