Course Schedule

Weekwise Schedule

Tentative and subject to change

Theme	Week	Topics
LLM Foundations I	1.1	Orientation, Transformer architecture
	1.2	Transformer Architecture - GPT 1 and 2
	2.1	Tokenization, Pretraining objectives
	2.2	Mixture of Experts
LLM Foundation II	3.1	Case studies: State-of-the-art open-source LLM architectures
	3.2	Scaling Laws, Emergent properties
GPU Basics	4.1	GPU architecture deep dive
	4.2	Parallelism: Multi GPU, Multi Node
	5.1	On-Prem Hardware Stack Deep Dive
Inference	5.2	Inference Strategies
	6.1	Inference Math and Bottlenecks
	6.2	Efficient Attention & KV Caching
Efficient Inference & Quantization	7.1	Quantization Fundamentals
	7.2	Inference Engines and Multi GPU
Fine-Tuning Fundamentals	8.1	Full Fine-Tuning vs. PEFT — When to Use Each
	8.2	Instruction Tuning
	9.1	Alignment (RLHF, DPO etc)
	9.2	More RL
Reasoning	10.1	Reasoning & Chain-of-Thought
	10.2	CoT, Tree-of-Thought, Self-Consistency — Prompt Engineering as Code
RAG	11.1	RAG Fundamentals - Context-engineering, embeddings, search and rerankers
	11.2	Evaluating RAG
Agents	12.1	ReAct Framework: Thought → Action → Observation
Tool Use & Function Calling	12.2	MCP introduction
	12.3	Agentic RAG, Multi Agent Orchestration, Multimodal Agents
Agent Finetuning	13.1	Fine Tuning for Tool calling
	13.2	Agent Evaluation & Safety
Evaluation	14.1	Evaluation
	14.2	Observability & Monitoring
Multimodal Models	15.1	Multi Modal Architecture: Image, Audio and Video models, Running Locally
	15.2	Fine tuning multimodal models
LLMs on the Edge	16.1	Edge-Optimized LLM Architectures, case studies
	16.2	Edge Optimization techniques
Security & Privacy Engineering	17.1	Threat Model: Prompt Injection, Jailbreaking, Data Leakage
Frontiers	17.2	Emerging Topics: Mamba, Qwen Next, Hybrid architectures
Presentations	18.1	Student Presentations I
	18.2	Student Presentations II

Materials

• Lecture slides and notes will be shared here as class progresses.