Learn to Build Your Own AI Tools: Lessons from Creating a Personal Banking App

Join us in our next Grenoble Data Science meetup.

Batiste will share his experience building GenAI tools such as a personal banking advisor.

Looking forward to seeing you!

Abstract
Curious about building your own generative AI tools but unsure where to start? In this talk, you’ll learn the practical steps to create AI systems for anything you need—drawing from my experience building a fully AI-driven personal banking advisor. Batiste believes anyone can craft their own GenAI tools—let’s explore how.
This talk will show you how to start from scratch and avoid some common pitfalls. Beyond the basics, Batiste will share personal lessons learned during his journey, hoping these can help you get started, and build tools better, faster, ... easier..
Here are some of the details you will gain by attending :
- How to add knowledge to an AI (RAG, web search, ...), while managing the context window well
- How to protect an AI from attacks (Guardrails): the multi-SLM approach
- How to make a basic front end for my AI (nicer than the ^^ terminal)
- Some tools that make my life easier (Arcade, LangSmith, WindSurf, ...)
- How do you get AI to make accurate calculations, even with small LLMs?
- How to make the AI remember what it is told
- How to reconcile reliability, costs, and response time
- When is it better to have a lot of small agents (or llm), or a big one?

Short Bio
Batiste Roger has spent the last months immersed in generative AI, creating systems for personal and professional purposes, such as a personal banking advisor and tools for generating song lyrics. He'd like to encourage anyone to make AI tools for whatever they care about, in an approachable, rewarding, and enjoyable way.

Special mention
We address a special thank you to the Cowork In Grenoble and Minimistan teams for their continued support in hosting us.