Skip to main content
Univelcity
EngineeringAdvancedNew

Advanced AI Agents

The frontier of AI isn't a smarter chatbot — it's systems that plan, use tools, and get work done with less supervision. Building agents that are both capable and controllable is one of the hardest and most valuable skills in AI right now.

Who it's for

  • AI and software engineers ready to build autonomous, tool-using systems.
  • Graduates of AI Engineering who want to go to the frontier.
  • Technical builders automating real, multi-step work.

Prerequisite: you've built with LLMs before — AI Engineering or equivalent experience.

What you'll build

A deployed multi-step (or multi-agent) system that automates a real workflow, with the evaluation and guardrails that make it trustworthy.

What you'll learn

  • Design and build agents that plan, reason, and use tools to complete real tasks.
  • Orchestrate multiple agents that collaborate reliably.
  • Give agents memory, control their behaviour, and keep them safe and predictable.
  • Evaluate and debug agentic systems — the hardest part of the job.
  • Deploy an agent that automates meaningful work.
  1. 01
    What an agent really is
    From single-shot prompts to planning, reasoning, and action loops.
  2. 02
    Tools and environments
    Giving agents capabilities, and designing the tools they use well.
  3. 03
    Planning and reasoning
    Task decomposition, reflection, and getting agents to think before they act.
  4. 04
    Memory and state
    Short- and long-term memory, context management, and continuity.
  5. 05
    Multi-agent orchestration
    Coordinating agents, roles, and hand-offs without chaos.
  6. 06
    Control, safety, and guardrails
    Keeping autonomous systems predictable and within bounds.
  7. 07
    Evaluating and debugging agents
    Observability and evaluation for systems that don't behave the same way twice.
  8. 08
    Ship an agent
    Deploy an agent that does real work. Capstone.
FORMAT

Live & hands-on

Live, expert-led, small cohort · 8 weeks · online · hands-on.
WALK AWAY

Certificate + proof of work

Univelcity Advanced AI Agents certificate + a live agentic system and repo.
TAUGHT BY

Practitioners

TODO — instructor details.

Questions

Good to know.

Is this for beginners?
No. It assumes you've already built LLM applications. Take AI Engineering first if that's not you.
What will I build?
A working agent, or multi-agent system, that automates a real task — deployed and evaluated.
How is this different from AI Engineering?
AI Engineering teaches you to build reliable LLM apps. This teaches you to build systems that act with autonomy — the harder, newer discipline on top.

Build agents that do the work.

Apply for the next cohort.