What I Learned from Google & Kaggle’s Generative AI Intensive Course

/ Hans van 't Riet

Last week, I joined over 100,000 participants in a 5-day Generative AI Intensive Course hosted by Google and Kaggle—a free and fast-paced program designed to equip professionals with practical knowledge on how to harness the power of GenAI in real-world settings.

Why did I join? Because GenAI is no longer a concept—it’s here, and it’s evolving faster than most organizations can absorb. As leaders in digital transformation, we can’t afford to wait. We need to understand not just the what, but also the how of these technologies.

This course offered an excellent foundation of the current status of GenAI technologies, how they can be applied today, and even provided glimpses into where they are likely to evolve next.

Below is a summary of the course—structured for executives and transformation leaders seeking clarity on how GenAI will impact their businesses.

Day 1: Foundational Large Language Models & Text Generation

Why it matters: Understanding the fundamentals is critical before scaling GenAI use cases. Day one unpacked the Transformer architecture, the core engine behind tools like ChatGPT and Gemini.

Key Takeaways:

  • LLMs are the brains behind GenAI—they interpret and generate human-like language at scale.
  • Transformer models help these systems understand context and nuance.
  • Fine-tuning allows you to adapt general models to business-specific tasks, such as customer service or marketing.

Google whitepaper: “Foundational Large Language Models & Text Generation”

Day 2: Embeddings and Vector Stores

Why it matters: Without intelligent data structuring, GenAI becomes just another flashy tool. This session focused on how to make AI actually useful inside your organization.

Key Takeaways:

  • Embeddings turn complex data into searchable formats.
  • Vector stores make this information retrievable at speed and scale.
  • Retrieval-Augmented Generation (RAG) combines LLMs with your proprietary data for smarter, context-rich answers.

Google whitepaper: “Embeddings & Vector Stores”

Day 3: Generative AI Agents

Why it matters: GenAI is moving beyond chatbots—into agents that can autonomously perform tasks, interact with systems, and even make decisions.

Key Takeaways:

  • AI agents integrate tools, logic, and memory to act independently.
  • Platforms like LangChain and Vertex AI Agents provide orchestration layers for real-world applications.
  • Think of these as junior digital employees—capable of assisting operations, support, or analysis at scale.

Google whitepapers: “Agents” and “Agents Companion”

Day 4: Solving Domain-Specific Problems Using LLMs

Why it matters: Generic models only take you so far. Tailoring AI to your industry delivers far more strategic value.

Key Takeaways:

  • Domain-specific LLMs adapt to unique challenges in sectors like healthcare and cybersecurity.
  • SecLM enhances threat detection and response capabilities in cybersecurity.
  • MedLM supports clinical workflows and patient information retrieval in healthcare.

Google whitepaper: “Solving Domain-Specific Problems Using LLMs”

Day 5: Operationalizing GenAI on Vertex AI with MLOps

Why it matters: Scaling GenAI requires more than a good prompt—it demands structured deployment, governance, and monitoring.

Key Takeaways:

  • MLOps for GenAI adapts best practices from machine learning to this new frontier of GenAI applications.
  • Understanding the GenAI lifecycle—from experimentation to production—is key to long-term success.
  • Platforms like Vertex AI help organizations deploy and manage GenAI responsibly and at scale.

Google whitepaper: “Operationalizing Generative AI on Vertex AI using MLOps”

My Reflections

This course reinforced a simple truth: GenAI is becoming more capable rapidly. And like any capability, it needs strategy, structure, and experimentation to create real business value.

If you’re in a leadership role, here are three questions to reflect on:

  1. Where can GenAI complement or augment your current operations?
  2. Do you have the data foundation to make it effective?
  3. Are you equipping your teams to experiment safely and learn quickly?

What’s Next

I’ll be diving deeper into some of these topics in future articles —especially GenAI agents and domain-specific applications.

Let’s continue learning and leading—together

Leave a comment