AI

servicios de IA para empresas | IBM , Automatizaciones personalizadas y formación en IA

AI courses for companies: agents, local LLMs and security

We wrote these courses ourselves for teams that will build, use or secure AI on their own infrastructure. Most of them end with something working: an agent that solves a real task, a model served on your GPU, or a set of security tests you can run again.

The ideal path

For a technical team that will run AI on its own servers, the order we recommend is: local LLMs and GPU infrastructure first, to choose the model, serve it with vLLM and measure how much load it takes before buying hardware; then LangChain for RAG and agents running locally, so the model answers from your documents; and finally AI agents, so it can also act on your systems in a real task. That is eight days, which can be spread over several weeks.

If the team already has an assistant or an agent running, the next step is the AI security course.

Or, if you prefer to go by level

Beginner. Copilot and ChatGPT training for businesses, for office and business teams already using an assistant.

Advanced. The AI agents course, which assumes software development experience but none in machine learning; the local LLM course, for systems teams; and, for those who already code in Python, deep learning with PyTorch and TensorFlow, LangChain for RAG and agents, and big data with Spark.

Expert. MLOps with Ubuntu Linux, when it is time to build the infrastructure to train and serve models, and AI security with OWASP, for security teams that already have an assistant or an agent running.

AI agents

Local LLMs

Deep learning, RAG and data

AI security

AI governance

Copilot and ChatGPT

Automation with n8n

If what you want is to automate processes with workflows and agents without writing an application, there are n8n basic and n8n advanced, one day each. The Zapier courses are still in the list below.

For the infrastructure behind your models

To build the platform where models are trained and served there is MLOps with Ubuntu Linux (UBU610), 3 days, and if you would rather we built it for you, our AI inference service on your own servers. If you work with watsonx, Watson or Cloud Pak for AIOps, IBM’s official training is under IBM AI.

How they are delivered

At your offices or remotely, with a live instructor. On the agents course, each person does the labs with the tool they already use, whether that is Copilot, Claude or another.

Tell us about your team’s profile and what you want to do with AI, and we will tell you where to start and which courses can be combined in the same week. For a complete role-based path, from infrastructure to AI governance, or a tailored programme, see AI training for companies.