Threads
Ideas that run through this notebook, with the modules that cover them.
Agentic AI
Two passes on the same topic: metrics first, then benchmarking pipelines.
MLOps lifecycle in one module, scaling and CI/CD in the other.
Comparison table and selection guide, then a tour of each tool.
Techniques, planning workflows and the patterns that package them.
Memory types and architecture, revisited with a memory manager.
Agent types, team architectures and collaboration patterns.
Threats and controls, then open challenges and limitations.
AI System Design
From a plain LLM app to RAG to agents.
Data, training, serving and the MLOps loop that ties them together.
Staying up, seeing inside the system and running it day to day.
Success metrics, evaluation and testing, then evaluation inside a full project.
Component-level access control and the wider security and privacy view.
Case studies, then a worked end-to-end project.
LangChain
From the big picture to a working install.
Pick a model, swap providers, shape the prompt.
Fixed pipelines, from a basic chain to RAG chains.
Tools first, then the agents that use them.
Load, split, embed, store, retrieve, answer.