Software Generation with Artificial Intelligence

Intensive online course (taught in English or Spanish)
20 hours · 2 weeks · Monday to Thursday · 7:00 PM – 9:30 PM (CET)
Live sessions + recordings available.

Learn how to integrate AI into the software lifecycle to speed up development, reduce errors, and improve decision-making.

70% hands-on / 30% theory
Support included
Access to recordings

Why this course?

Technical teams that understand how to integrate AI into their workflow build software faster, with fewer errors, and with better-informed decisions.

Immediate impact

Exercises applied to real-world scenarios in development, QA, security, and data.

Practical methodology

70% live practice with ready-to-use repositories and tools.

Measurable results

Time savings, fewer bugs, and improved deliverable quality.

Who is it for?

What you will learn?

Create coding assistants aligned with internal policies

Generate and refactor code with AI

Detect vulnerabilities from the start

Build lightweight apps based on open models

Combine RAG + MCP for secure and traceable workflows

Best practices and responsible evaluation

70% practical, 30% theory. Entirely online, with recordings available.

Schedule

Requirements

Learn from Artificial Intelligence specialistsa

Zuzanna Parcheta is a software engineer and specialist in artificial intelligence applied to development and process automation. She has worked on international projects with high technical demands, including her time in Luxembourg at the Translation Centre, where she contributed to the development of machine translation systems that outperformed commercial engines and even those of the European Commission, combining advanced NLP, deep learning, and large-scale production deployment.

With a strong focus on real-world impact, Zuzanna helps technical teams integrate AI into their daily workflow — generating and reviewing code with higher quality in less time, automating documentation, creating tests and security validations, and accelerating product delivery without losing technical control.

Beyond building solutions, Zuzanna trains other engineers, analysts, and technical leads to use AI professionally (not just “try prompts”). Her approach is practical and transparent: real-world cases, step-by-step guided exercises, and tools ready to apply from day one in your own projects.

Ready to learn with Zuzanna?
Reserve your spot and start transforming the way you build software with Artificial Intelligence.

What are the modules you’ll see in the course?

Module 1 — AI Across the Software Lifecycle

What you’ll learn:
How AI augments each SDLC stage: requirements, design, implementation, testing, deployment, maintenance.
Realistic benefits and common risks (quality, speed, privacy, over-reliance).
Hands-on:
Short prompts that transform routine tasks (explanations, refactors, tests).

Module 2 — Prompting That Delivers

What you’ll learn:
Good vs. bad prompts; briefing an AI like a junior developer; structure, constraints, and examples.
Creating a lightweight custom GPT aligned to coding guidelines.
Hands-on:
Build and test a simple, policy-aware coding assistant; compare outputs across models with a checklist for correctness and style.

Module 3 — Code Generation in Practice

What you’ll learn:
How LLM code generation works (context, instruction following, fine-tuning).
Tooling options in editors and the web.
Hands-on :
Generate functions and a small app from a natural-language spec; iterate via feedback instead of manual hot-fixing.

Module 4 — Refactoring with AI

What you’ll learn:
Refactoring vs. optimization; “measure before you cut”; readability, complexity, tests, security, performance baselines.
Free tooling for code quality, testing, security, profiling across languages.
Hands-on:
Clean up “spaghetti” code; compare complexity and run time before/after; justify changes with metrics.

Module 5 — Optimizing Computational Cost

What you’ll learn:
Time, memory, and money: tying performance to compute cost; vectorization and algorithmic choices.
Hands-on:
Speed up a nearest-neighbor baseline using NumPy (and optional KD/BallTree or Numba); verify identical outputs and report speedups and cost deltas.

Module 6 — From Model to Chat App

What you’ll learn:
Building blocks of a GenAI program: tokenizer, transformer, decoding, serving, and UI options (CLI/Gradio/FastAPI).
Hands-on:
Run a small open model locally or on Colab, then build a minimal chat interface and test simple tasks end-to-end.

Module 7 — AI-Assisted Vulnerability Detection

What you’ll learn:
SAST/DAST basics, SBOMs, CWE/CVE mapping; where AI helps and its limits; secure review workflows in CI.
Hands-on:
Break and fix a tiny Flask app using Bandit/Semgrep/pip-audit; add an AI review step with precise, minimal diffs and reasoning.

Module 8 — MCP Servers: Let Models Act Safely

What you’ll learn:
What MCP is and isn’t; tools, resources, prompts; stdio/SSE transports; least-privilege, validation, and audit trails.
Hands-on:
Build a small MCP server in Python; expose one tool and one resource; explore it with MCP Inspector and a CLI.

Module 9 — LangChain & RAG: Let Models Know Safely

What you’ll learn:
Core LangChain concepts (models, prompts, parsers, chains/LCEL, tools/agents, retrievers/ RAG) and when to use them.
Chunking, metadata, embeddings, vector stores; structured outputs and guardrails.
Hands-on:
Lab A/B/C: scaffold a repo from a spec via RAG, generate grounded README/tasks with citations, and run an approval-gated agent with file tools.

Module 10 — “Grounded Action” Capstone: Combine RAG + MCP

What you’ll learn:
Decision rules: when to prefer MCP (act) vs. RAG (know), and why the best pattern is often retrieve → decide → act → explain.
Hands-on:
Implement a small workflow: retrieve a policy (RAG), execute a safe action through an MCP tool, and produce a cited summary for auditability.

Module 11 — Trustworthy & Explainable AI

What you’ll learn:
Core principles of Trustworthy AI: fairness, accountability, transparency, and human oversight. Explainability vs. interpretability — why both matter for compliance (AI Act, ISO 42001, NIS2).
Bias, robustness, and data governance in AI-assisted software systems.
How to integrate explainability into development pipelines and documentation.
Hands-on:
Use open-source libraries (e.g., SHAP, LIME, ELI5) to generate model explanations. Design a mini “transparency card” summarizing model purpose, data sources, and limits for an internal AI tool.

Pricing

Companies

Team package up to 15 participants → €5.499 per edition
Flexible scheduling
Dedicated Q&A space for the team

Individuals / Freelancers

Individual price → €449 (VAT included)
Access to recordings
Certificate of completion
Email support
Next edition: January 12 2026

Frequently Asked Questions

In what language is it taught?

In English or Spanish. The support team will answer your questions in your preferred language.

Will I have access to recordings?

Yes, all sessions are recorded and provided to the students.

Do I need previous experience?

We recommend basic knowledge of programming/scripting and familiarity with Git.

How do I register?

Fill out the form and we’ll contact you to confirm your spot. If you choose the company format, dates will be scheduled by arrangement; if you choose the individual format, the next edition starts on January 12, 2026.

Contact

Registration

Nombre

Apellidos

Correo electrónico

Idioma del curso

Modalidad