Agentic AI Bootcamp

• Focus: Build your first LLM chain with minimal setup
  
• Key Concepts: API key management, Groq, ChatOpenAI, PromptTemplate, LLMChain, ConversationBufferMemory
  
• Hands‑On: Simple FAQ chatbot that remembers context

   

Case Study: Product‑support FAQ assistant

• Focus: Add long‑term context and ground answers with documents
  
• Key Concepts: SummaryMemory, VectorStoreRetrieverMemory, embeddings + FAISS, RAG workflow fundamentals
  
• Hands‑On: Extend the FAQ bot to a mini‑RAG over docs

   

Case Study: Internal knowledge‑base assistant for an SME

• Focus: Foundations of agent reasoning & tool use
  
• Key Concepts: ReAct, Plan‑and‑Execute, function‑calling, tool‑design patterns
  
• Hands‑On: ReAct agent that plans a weekend trip (flight API stub, weather, maps)

   

Case Study: Kayak‑style concierge

• Focus: Standardised tool & memory interfaces with MCP
  
• Key Concepts: MCP server anatomy (stdio, HTTP‑SSE), list_tools(), wrapping MCP tools in LangChain
  

Hands‑On: Spin up mcp‑sandbox, expose calculator & weather tools, call them from an LLM chain

• Focus: Multi‑step planning & observability (MCP‑aware)
  
• Key Concepts: initialize_agent types, tool routing & fallback, LangSmith tracing
  
• Hands‑On: Upgrade Session 3 trip planner with LangSmith traces

   

Case Study: Expedia‑style itinerary generator

• Focus: Build your first CrewAI multi‑agent crew
  
• Key Concepts: Agents, Tasks, Crews; short‑ vs long‑term memory; flow‑state export; event emitters
  
• Hands‑On: “Researcher + Writer” crew that drafts a blog post

   

Case Study: Automated article writer

• Focus: Production‑quality CrewAI workflows
  
• Key Concepts: Conversation Crew, multimodal abilities, HITL loops, knowledge‑source tools
  
• Hands‑On: Customer‑support crew with HITL escalation

   

Case Study: SaaS support bot

• Focus: Graph‑based agent orchestration, branching, streaming, fault‑tolerance
  
• Key Concepts: Nodes, edges, state; create_react_agent; branch/merge paths; retries; streaming through state
  
• Hands‑On: DAG chaining extraction → enrichment → summarisation agents

   

Case Study: Media‑monitor news summariser

• Focus: Vector DB integration & structured output
  
• Key Concepts: Embedders (HF, VoyageAI), vector DBs (Chroma, Qdrant); file loaders; JSON‑schema outputs; memory in CrewAI & LangGraph
  
• Hands‑On: Legal‑doc analyst agent storing rulings in Qdrant with citations

   

Case Study: Legal case‑law assistant

• Focus: Testing and securing agent behaviour

   

• Key Concepts: LangSmith eval suites, LangFuse JSON logs, prompt‑layer testing, guardrails with Guidance / JSON schema

   

• Hands‑On: Compare two prompt strategies for the analyst agent and visualise results in LangSmith

   

Case Study: E‑commerce gift‑advisor accuracy audit

• Focus: Exposing agent APIs for production

   

• Key Concepts: FastAPI streaming endpoints, authentication, rate‑limits, secrets management, API versioning, caching

   

• Hands‑On: Wrap an agent in FastAPI, deploy behind an API gateway

   

Case Study: SaaS agent API gateway

• Focus: Containerisation and serverless scaling

   

• Key Concepts: Docker multi‑stage builds, image security, orchestration basics, Cloud Run deployment & autoscaling, health checks, logging

   

• Hands‑On: Containerise the agent and an MCP server, push to Artifact Registry, deploy to Cloud Run

   

Case Study: Multi‑tenant agent SaaS