Agent Dev Academy Request a seat

Live, hands-on agent engineering

Build agents that do the work.

Learn to design, build and run AI agents with Claude, OpenAI and open-source tools, guided by experienced tutors. Bring a real workflow and build an agent that runs it.

Request a seat See the curriculum

Curriculum

From first loop to production.

Each module ends with working code in your own repo, built around a workflow you actually want automated.

  1. How agents really work

    The loop behind every agent: model, tools, memory and stopping rules. When an agent is the right call, and when a plain script is better.

  2. Tools and integrations

    Connect agents to email, docs, CRMs, databases and internal APIs with function calling and MCP, with permissions you control.

  3. Build with coding agents

    Use Claude Code and similar tools to write, test and refactor faster, while you review and own every change.

  4. Evals, guardrails and shipping

    Test agents like software, cap cost and access, keep a human on the risky steps, and monitor them once they run on their own.

Format

Build it, test it, ship it.

Live

Live, not recorded

Learn in live sessions with experienced tutors. Bring your code and your questions.

Project

Your workflow, your agent

Work on a real process from your job, not a toy demo. The goal is an agent you keep using.

Production

Production-ready habits

Version control, evals, logging and cost limits: the way teams run agents for real.

Who it's for

For people who build, not just prompt.

Some coding helps: if you can read a script and use a terminal, you're ready. Bring a laptop and a workflow you want automated.

  • Developers

    Software engineers

    Who want to add agents to their products and internal tools.

  • Product

    Technical PMs and founders

    Who need to know what agents can really do before betting a roadmap on them.

  • Operations

    Ops and automation leads

    Who already run scripts and no-code flows, and want them to handle the judgement calls too.

  • Data

    Analysts and data people

    Comfortable with SQL or Python, and ready to automate the analysis itself.