Most developers use LLMs through prompts, but production AI systems require greater control, structure, reliability, and scalability.

In this session with hands-on demo, we will explore how to build real-world AI applications using Gemma, an open-weight model. Participants will learn the full lifecycle of working with modern language models—from tokenization and decoding strategies to structured outputs and tool integration through function calling.

What you will learn:

• How tokenization and decoding influence model behavior and output quality

• Generating structured JSON outputs for reliable automation

• Connecting Gemma with APIs, tools, and external systems

• Building an end-to-end AI workflow beyond chat interfaces

• Best practices for production-ready AI systems

This session is ideal for developers, software engineers, AI practitioners, and anyone looking to move from prompt experimentation to building scalable AI products.

By the end of the session, participants will have practical knowledge and a working prototype for creating controllable, composable AI systems.

Note: Please bring your laptop for hands-on demo .