ՀԱՅEN
welcome.txt

Hi! I'm Arman Ayvazyan. Welcome to my space!

Here I share what I learn about AI in QA and test automation, and run practical online lessons to help you upgrade your technical skills.

What you'll find here

Courses

02 · Agents

Building Agentic Workflows

Design agents that plan, use tools and finish real tasks safely.

  • 12 h · 6 × 2 h
  • Live online · Recording

What you'll be able to do

  • Decide when a task needs an agent and when a simple script is enough
  • Build an agent loop with tools, memory and clear stop conditions
  • Split work across several agents with a shared contract
  • Add guardrails: permissions, human approval and spending limits
  • Test and observe agents so you can trust them in production

Program

  1. Agents vs workflows

    Patterns: chains, routing, parallel work, orchestrator and workers.

  2. Tools and tool calling

    Design tools the model can use well, and handle their errors.

  3. Memory and context

    What to keep, what to summarise and what to forget.

  4. Multi-agent systems

    Planner, workers and reviewers working from one shared plan.

  5. Safety and control

    Permissions, approvals, budgets and prompt-injection defence.

  6. Evaluate and ship

    Evals, traces and logs, then deploy your agent and keep it healthy.

Who it's for

Developers and QA engineers who already use an AI assistant and want to automate whole tasks. Basic programming (JavaScript or Python) required.