New training program

Agentic Engineering Course

The discipline of building agentic systems that perform with institutional precision.

For teams who understand that AI is not used — it is engineered. Over four weeks we instill the rigor required to design autonomous systems: from precise problem modeling to multi-agent orchestration that delivers measurable, defensible value.

This is not a course of isolated prompts.

This is not an experimentation workshop.

It is engineering training to build production-ready agentic systems.

I want to bring this program to my team View full program

Objective

Program objective

We want the team to stop using AI reactively and start:

modeling problems as agentic architectures

designing professional contexts and skills

building real, useful tools

implementing agents with memory and planning

orchestrating multi-agent systems

integrating this capability into actual development processes

Real problem

Your team already uses AI. The problem is it does so without engineering.

Most technical teams already use ChatGPT, Claude, Cursor, or similar tools. But they do it in isolation, without architecture, without observability, and without a shared methodology.

unpredictable and hard-to-reproduce results

dependence on a few people who 'figured it out'

lots of trial and error, little structural progress in how software is built

This program fixes that with real engineering rigor.

What the team learns

From using AI, to engineering with AI

This is not a prompts course. It is applied training so your team learns to:

deeply understand how LLMs work and what their real limits are

perform discovery and specification using AI as a design tool

use code assistants with engineering judgment

design robust contexts, skills, and persistent instructions

build agents equipped with real tools

incorporate memory, planning, and patterns like ReAct

orchestrate collaboration between multiple agents

bring all of this into daily development work

Outcomes

What the team takes away in 4 weeks

a solid mental model for designing agentic systems

professional judgment for choosing and combining tools

an applicable methodology for agentic engineering

a reusable base kit of contexts and skills

a functional agent with tools built during the course

an evolved version of the agent with memory and/or planning

practical introduction to multi-agent orchestration

a concrete adoption plan for the team

Detailed program

Program

Week 1

Foundations + Agentic Modeling

Session 1

LLMs and reasoning for agentic systems

Establish a solid technical foundation on how models actually work and what it means to design systems around them.

What an LLM is and what real capabilities it has

Tokens, context windows, and practical constraints

Pretraining, fine-tuning, and inference without marketing

Why agents need structure, not just prompts

Hallucinations, context drift, and how to mitigate them

How to choose models based on their role in a system

Session 2

Agentic discovery and professional specification

Use AI to do better analysis and design work, not just generate code.

Professional use of chat interfaces for discovery

Turning conversations into structured specifications

Modeling a problem as an agent flow

Defining responsibilities, tools, and handoffs

Documentation that actually helps building

Prompt patterns for requirements and architecture

Week 2

Context Engineering & Agentic Workflows

Session 3

Context engineering and agentic development tools

Learn to build contexts and workflows that produce reliable, maintainable results.

Context engineering vs casual prompting

Structure for agentic projects (AGENTS.md, skills, etc.)

Professional code assistants and how to guide them

When to accelerate with AI and when to keep strong human control

Patterns for generating code, tests, migrations, and docs

Controlled iteration and technical review

Session 4

Skills, operational memory, and professional setup

Teach the team how to create the right environment so agents can operate reliably and repeatably.

Designing reusable skills

Base file structure and persistent prompts

Managing long context and compression

Difference between working memory and project memory

Professional initialization of agentic projects

Versioning and collaboration on contexts

Week 3

Building Agents

Session 5

Building a functional agent with tools

Build the first agent that actually does real work.

Basic architecture of a tool-using agent

Designing and implementing useful tools

Reasoning and execution loop

Connecting to models and error handling

Internal vs external tools (APIs, databases, etc.)

Behavior evaluation and basic testing of agents

Guided construction of the first real agent

Session 6

Advanced agents: memory, planning, and robustness

Evolve the agent into something more autonomous and reliable.

Designing short-term and long-term memory

Task planning and decomposition

ReAct patterns in practice

Self-correction and reflection

Techniques for maintaining quality in long sessions

Basic observability and agent debugging

Week 4

Multi-Agent Orchestration & Adoption

Session 7

Multi-agent systems: design and orchestration

Understand when and how to compose multiple agents to solve more complex problems.

Single agent vs team of agents

Orchestration patterns: supervisor, planner, specialists

Feedback and handoff flows

Recovery from failures

Designing collaboration protocols

Risks, limits, and good design practices

Session 8

Final project and organizational adoption

Consolidate everything learned and define how to bring it into the team's daily reality.

Presentation and review of final projects

How to introduce agents in discovery, development, review, and QA

Realistic integration with existing workflows

Light governance, human oversight, and accountability

Metrics that actually matter

Concrete adoption plan for the following weeks

Final project

Throughout the program, each participant develops a project that demonstrates mastery of agentic engineering:

clear modeling of the problem as an agentic system

professional design of skills and context

use of code assistance tools

a functional agent equipped with tools

evolution of the agent with memory and/or planning

a simple orchestration or collaboration scheme between agents

The goal is to finish with something that can serve as a real foundation for the team.

Methodology

The methodology combines:

live classes with demonstrations

step-by-step guided construction

applied exercises

progressive project

peer review and feedback

Understand → Design → Build → Evaluate → Adopt

Benefits for the company

significantly accelerates design and prototyping work

reduces friction to move from ideas to implementations

installs a shared language and practice around AI

reduces chaotic experimentation

generates reusable assets (contexts, skills, agents)

prepares the team for increasingly agentic development

Benefits for the team

truly understand how LLMs work and how to use them well

gain technical judgment for choosing and combining tools

build real agents and evolve them

learn to work with context, memory, and tools professionally

develop a professional agentic engineering methodology

leave with concrete capability to apply what was learned

Requirements

Recommended requirements

prior software development experience

understanding of architecture and APIs

comfort working with code, repositories, and modern tools

willingness to build during the program

Why Stock42

Stock42 does not teach this from theory or hype. It teaches it from real practice.

2 years designing and operating agentic systems in production

more than 50 projects developed with agentic architecture

real experience in products, automation, and technical adoption

focus on results for startups, SMBs, and teams that need engineering, not fashion

Investment

Up to 5 participants: US$ 2,490

Up to 10 participants: US$ 3,990

Up to 20 participants: US$ 5,990

Additional participant: US$ 290

Ideal for startups, SMBs, and technical teams that want to incorporate AI seriously and at scale.

Request proposal

FAQ

Do we need prior experience with agents?

No. The program is designed to take the team from solid fundamentals to building real systems.

Is it a technical or strategic course?

It is strongly technical with practical application. It is designed so the team leaves building, not just understanding concepts.

Is it useful for seniors who still do not master LLMs?

Yes. It is especially valuable for profiles with solid technical experience who want to incorporate AI in a rigorous way.

Is the course generic or adapted to the team?

It is delivered with an in-company focus and can be contextualized based on stack, product type, and team maturity.

What happens after the course?

The team leaves with methodology, concrete projects, base contexts, and a clear plan to continue incorporating agentic systems into daily work.

Your team does not need to use more AI. It needs to engineer with it.

This program is designed for teams that want to stop being spectators of change and become the ones building the new way of developing software. The goal is not to replace engineers. The goal is to multiply their capacity in a professional way.

I want to talk to Stock42