Skip to main content
Ace Skills Consulting
Advanced

Agentic AI with LangChain

Design, build and evaluate reliable AI agents with LangChain and LangGraph — from a single tool call to multi-agent workflows.

  • 24 hours of live, trainer-led sessions
  • Build agents with LangChain and LangGraph
  • Agentic RAG, memory and human-in-the-loop steps
  • Evaluation, tracing and guardrails for production
  • Capstone: a multi-step agent for a business workflow
Duration
3 Days | 24 Hrs
Next batch
Batches on request
Delivery
Live Online · Onsite · Corporate
Level
Advanced
Language
English

Why learn this course with Ace Skills

  • Agents as engineering

    Model agents as explicit graphs so behaviour is predictable, testable and debuggable.

  • Multi-step workflows

    Combine tools, retrieval, memory and approvals into end-to-end automations.

  • Measure reliability

    Build evaluation sets and traces to prove an agent works before you ship it.

  • Guardrails built in

    Defend against prompt injection and limit what agents are allowed to do.

Course overview

AI agents go beyond answering questions: they plan, call tools, read data and take actions across several steps. Building agents that behave reliably is an engineering discipline, and this three-day course teaches it hands-on using LangChain and LangGraph.

Day one establishes the building blocks: chat models, prompts and structured output with LangChain, tool calling, and the reasoning-and-acting loop. You will then move to LangGraph to model agents as explicit graphs of state, nodes and edges, which makes control flow, retries and branching far easier to reason about.

Day two adds knowledge and memory. You will build agentic retrieval-augmented generation that decides when and how to search, add short- and long-term memory, persist state with checkpoints and include human approval before sensitive actions. You will also connect agents to external tools and data, including the Model Context Protocol.

Day three focuses on reliability: multi-agent patterns such as supervisors and specialists, tracing and debugging, building evaluation datasets, measuring task success and cost, and adding guardrails against prompt injection and unsafe actions. The course ends with a capstone in which you build, evaluate and demonstrate an agent that automates a realistic business workflow.

This is an original Ace Skills course with no vendor exam. On completion you receive an Ace Skills Consulting course completion certificate.

Key features

  • LangChain and LangGraph from fundamentals to multi-agent systems
  • Tool calling, agentic RAG and memory
  • Human-in-the-loop approvals and checkpoints
  • Model Context Protocol integration
  • Tracing, evaluation and cost tracking
  • Guardrails and security patterns
  • Capstone agent project
  • Course completion certificate from Ace Skills Consulting

Who should attend

  • Python developers building LLM applications
  • Machine learning and data engineers
  • Solution architects designing AI systems
  • Generative AI for Professionals graduates with coding experience

Prerequisites

  • Solid Python programming skills
  • Experience calling LLM APIs and basic RAG concepts
  • Familiarity with JSON and REST APIs

Curriculum

3 Days · 24 Hours · 10 Modules · 6 Labs

  1. From LLM calls to agents

    • What makes an agent
    • Reasoning-and-acting loops
    • When not to use an agent
    • Cost, latency and reliability trade-offs

    LangChain building blocks

    • Chat models and prompt templates
    • Structured output with schemas
    • Defining tools
    • Tool calling
    • Lab: Build a tool-calling assistant for currency and calendar tasks · 45 min

    LangGraph fundamentals

    • State, nodes and edges
    • Conditional routing
    • Loops and stopping conditions
    • Error handling and retries
    • Lab: Rebuild the assistant as a LangGraph agent · 60 min

  2. Agentic RAG

    • Retrieval as a tool
    • Query rewriting
    • Self-checking answers against sources
    • Hybrid search
    • Lab: Agentic RAG over a product knowledge base · 60 min

    Memory and persistence

    • Short-term conversation state
    • Long-term memory stores
    • Checkpoints and resuming runs
    • Human-in-the-loop interrupts
    • Lab: Add memory and an approval step before actions · 45 min

    Connecting to the world

    • Model Context Protocol servers and clients
    • Calling internal APIs and databases
    • Streaming and async execution

  3. Multi-agent patterns

    • Supervisor and specialist agents
    • Hand-offs
    • Planning and plan-execute patterns
    • Shared state
    • Lab: Build a supervisor with two specialist agents · 45 min

    Evaluation and observability

    • Tracing runs
    • Evaluation datasets
    • Task success and trajectory evaluation
    • Tracking cost and latency

    Guardrails and deployment

    • Prompt injection defences
    • Tool permissions and least privilege
    • Output validation
    • Deploying agents as services

    Capstone

    • Designing the workflow
    • Building and evaluating
    • Demonstration and review
    • Lab: Build, evaluate and demo a business workflow agent · 120 min
Download full syllabus (PDF)

Skills you will gain

  • Agentic AI
  • LangChain
  • LangGraph
  • Tool calling
  • Agentic RAG
  • Agent memory
  • Human-in-the-loop
  • Multi-agent systems
  • Model Context Protocol
  • LLM evaluation
  • Tracing
  • AI guardrails

Tools & technologies

  • Python
  • LangChain
  • LangGraph
  • LangSmith
  • LLM APIs
  • Vector database
  • FastAPI
  • Jupyter Notebook

Schedule & Pricing

Live Online

Individual

Price on request

  • 24 hours of live, trainer-led sessions
  • 6 hands-on labs
  • Ace Skills course completion certificate

Onsite

At your office or our centre

Pricing depends on location, group size and dates.

  • Delivered at your premises or an Ace centre
  • Same curriculum and labs as live online
  • Dates planned around your team

Corporate / Team

Private batch for your organisation

A private batch with content adapted to your environment and goals.

  • Customised curriculum
  • Private batch, online or onsite
  • Progress and assessment reports
  • GST invoice and PO-based billing

No public batches scheduled — Talk to an Advisor for a private batch.

Projects & labs you will build

  • Day 1 · 60 min

    Rebuild the assistant as a LangGraph agent

  • Day 3 · 120 min

    Build, evaluate and demo a business workflow agent

Your certificate

Certificate of Completion — Ace Skills Consulting

Awarded on completing the course. It records the course title and training hours, and you can share it on your LinkedIn profile.

Frequently asked questions

Is there an exam?

No. This is an original Ace Skills course with no vendor exam; you receive an Ace Skills course completion certificate.

How much Python do I need?

You should be comfortable writing functions, classes and working with packages and APIs; this is an advanced course.

Which models do the labs use?

Labs are model-agnostic and work with several leading LLM providers; we discuss trade-offs between them.

Will the code work after the course?

Yes. You keep the notebooks and capstone code, and we note which library versions they were tested against.

Should I take Generative AI for Professionals first?

If you have not built a RAG application or called LLM APIs before, yes — it covers the groundwork this course assumes.

Related courses

Explore more

All Data & AI courses

Train your team on Agentic AI with LangChain

Private batches online or at your office in India and the GCC, with a curriculum adapted to your environment.