Course: Ai Agents Pro

Learn how to design, run, and scale AI agents using popular frameworks.

Mentor - Oleksandr Denisyuk, CTO @ Ukrposhta
Program
Course: Ai Agents Pro

AI Agents Pro is an advanced course on creating and implementing AI agents in real projects. You will get a systemic understanding of designing agent systems, master the modern 2026 stack: Claude Agent SDK, OpenAI Agents SDK, Google ADK, LangGraph and CrewAI. Multi-agent approaches, agent RAG, MCP/A2A protocols, security and guardrails, observability and evaluation, as well as learn how to deploy solutions in the cloud with the basics of AIOps.

We teach how to build agents the right way: with thoughtful architecture, security, tests and monitoring. You will delve into the development of production-ready solutions in practice.

Mentor - Oleksandr Denisyuk, CTO @Ukrposhta, AI Researcher, 25 years in the IT industry: from engineer to CTO, CEO.

What you will get after the course:

  • ✅ Deep understanding of agent architecture
    Single- and multi-agent systems, design patterns, context engineering, memory and human-in-the-loop.
  • ✅ Hands-on skills with a modern stack
    Agent SDK (Claude, OpenAI, Google ADK), LangGraph, CrewAI; the ability to choose a framework for a specific task.
  • ✅ Agent RAG and tool integration
    Chunking, vector DBs, hybrid search + reranking, GraphRAG; MCP, A2A, ACP protocols and creation of own MCP servers.
  • ✅ Security agents and guardrails
    Prompt injection, sandboxing, least-privilege for tools, Guardrails AI and NeMo Guardrails.
  • ✅ Observability and evaluation
    OpenTelemetry GenAI, Langfuse, LangSmith, Phoenix, cost control; Ragas, LLM-as-judge and promptfoo in CI.
  • ✅ Cloud deployment and AIOps basics
    AWS Bedrock AgentCore, GCP Vertex Agent Engine, Kubernetes, serverless, scaling and secure access.
  • ✅ Bonus: Spec-Driven Development (SDD)
    How specifications guide AI agents in the construction of Enterprise-level systems based on real practical examples.
  • ✅ Portfolio + case from the hackathon
    Production-ready PoC, which can be used as a start for a new product based on AI agents.

The course is equally valuable for developers (how to build agents) and for DevOps / SRE (how to safely operate, monitor and deploy them).

Format:

    Duration: 9 online lessons of ~2.5 hours (first lesson - 3 hours)

    Dates: 17.08 - 04.09, start at 18:30 (Kyiv time, GMT+2).

    Platform: Zoom. You will receive a link to the stream before the course begins, and it will also be available on this landing page. In addition, all materials and lecture recordings will be available on the educational platform.

    Interaction: lectures, meetings, practical and homework. Participants will be able to defend their project at the final Vibeathon and receive feedback from the jury, as well as gifts from the organizers!

    Language of the event: Ukrainian

    Language of presentations: Ukrainian with English terms

    *Basic knowledge of Python is recommended
    **We work with Github, so it is recommended to have a Github account.
    ***Before starting the course, you need to purchase a subscription to Cursor / Claude Code / Windsurf / Codex / Cline.

Program

17.08 (Monday, 3 hours) — Fundamentals of Agents 2026
  • Introduction to the course: structure, tools, tasks (15 min)
  • Theory: fundamentals of AI and LLM, agent loop and ReAct, single-agent systems and tools, 4 patterns of effective agents; agentic coding (Cursor / Claude Code) (1–1.5 h)
  • Questions and answers, break (10 min)
  • Workshop + lab: your first agent with tools (50 min)
19.08 (Wednesday, 2.5 h) — RAG → Agent RAG
  • Theory: RAG basics, chunking, vector DBs, RAG types, hybrid search + reranking, GraphRAG, agent RAG (Self-RAG, CRAG) (1 hour)
  • Questions and answers, break (10 min)
  • Workshop + laboratory: agent RAG on own documents (50 min)
21.08 (Friday, 2.5 hours) — Frameworks and Agent SDK Landscape
  • Theory: LangGraph vs Claude Agent SDK vs Google ADK vs CrewAI; when to choose what; workflows and middleware (1 hour)
  • Questions and answers, break (10 min)
  • Practice + laboratory: the same agent in two stacks (40 min)
24.08 (Monday, 2.5 hours) — Multiagent systems + Context Engineering
  • Theory: multiagent vs single-agent, patterns (orchestrator–worker, handoffs, routers, parallelization), context engineering, memory, human-in-the-loop (1 hour)
  • Questions and answers, break (10 min)
  • Practice + lab: multi-agent solution with memory (40 min)
26.08 (Wednesday, 2.5 hours) — Protocols: MCP, A2A, ACP
  • Theory: MCP deep-dive (servers, clients, authentication, security), A2A, ACP (40–60 min)
  • Questions and answers, break (10 min)
  • Practice + laboratory: developing and using our own MCP server (40 min)
08/28 (Friday, 2.5 h) — Agent Security and Guardrails
  • Theory: prompt injection, data exfiltration, tool misuse; isolation; Guardrails AI, NeMo Guardrails (40 min)
  • Q&A, break (10 min)
  • Practice, hands-on: setting up guardrails and protecting the agent (60 min)
31.08 (Monday)2.5 hours) — Observability and Evaluation
  • Theory: OpenTelemetry, tracing (Langfuse, LangSmith, Arize Phoenix), metrics; Ragas, LLM-as-judge, DeepEval (40 min)
  • Q&A, break (10 min)
  • Practice + lab: instrumenting an agent and building an eval pipeline (60 min)
02.09 (Wednesday, 2.5 h) — Production, deployment and AIOps basics
  • Theory: AWS Bedrock AgentCore, GCP Vertex Agent Engine; containers and Kubernetes; serverless, scaling, logging, tracing, secure access; AIOps / agent-ops basics (40 min)
  • Q&A, break (10 min)
  • Practice: deploying an agent to the cloud (30 min)
04.09 (Friday, 3 hours) — VIBEATHON
  • Vibeathon with analysis of participants' projects: design through SDD and build a production-ready PoC
  • Voting and awards

The course concept and all course materials are the intellectual property of FWDAYS. Copying of course content and materials is possible only with the consent of FWDAYS.

Who will be interested?

  • Developers
  • DevOps Engineers
  • System Administrators
  • QA
  • Technical Specialists with a Code Understanding


Mentor: Oleksandr Denisyuk, CTO @ Ukrposhta

— Senior R&D manager, AI Researcher

— 25 years in the IT industry: from engineer to CTO, CEO

— Startup founder, change manager

—Author of the TG channel AI no magic

—Host of the Youtube channel AI no magic

Linkedin

Event price

Attendee's ticket

Access to 9 online broadcasts (~20 hours)

Presentation and materials

Access to the course recording on the educational platform

Access to chat in TG, where the speaker will answer questions

Certificate of passing (provided homework is completed)

Free month at Fwdays Club

Course participants receive a 10% discount on participation in Fwdays conferences

Available payment in installments from Monobank and purchase in installments from Privatbank
If after the first lesson you realize that the expectations do not match, we will refund the money in full.
16 500 UAH ≈€330
till 6 August
till 18 June — 13600 UAH≈€272
till 19 August — 18900 UAH≈€378
Buy ticket
−10% for participants of previous conferences 
It will be added automatically to account where you made a purchase previously
−30% for students of full-time education 
Send us a photo or a scanned copy of your Student's ID to academy@fwdays.com and we will send a promo code to you
Діють спеціальні умови для військовослужбовців та УБД.
Заповніть форму, щоб отримати персональний промокод
Group discounts are available for companies
Напишіть нам на academy@fwdays.com для прорахування вашої знижки.
Sign in
Or by mail
Sign in
Or by mail
Register with email
Register with email
Forgot password?