Wolt x Doctolib AI Meetup

Tauche ein in die faszinierende Welt der intelligenten Systeme! Erlebe exklusive Einblicke von Experten, die zeigen, wie moderne Algorithmen den Alltag revolutionieren – vom automatisierten Einkauf bis hin zur effizienten Softwareentwicklung im Gesundheitssektor. Lerne aus erster Hand, wie man komplexe Herausforderungen meistert und innovative Lösungen in die Realität umsetzt. Nutze diese einzigartige Gelegenheit, um dich mit Gleichgesinnten zu vernetzen, wertvolles Wissen zu teilen und die Zukunft der Technologie aktiv mitzugestalten. Sei dabei, lass dich inspirieren und bring deine eigenen Projekte auf das nächste Level!

You're invited to the first edition of our AI meetup, hosted in collaboration with Doctolib. In this meetup, Wolt’s and Doctolib’s AI experts will share their experience, learnings and advice on building AI applications in the quick commerce and health domains. Wolt Agentic Shopping: The Agent Is Already in the Aisle by Yesid López, Senior ML Engineer @ Wolt What happens when a shopping assistant turns a messy grocery list, or even something vague like "something for dinner," into real baskets from nearby stores? This talk looks at an internal Wolt experiment: a Grocery Basket Builder designed to reduce the friction of grocery shopping by helping users find, compare, and adjust baskets across stores. Along the way, it explores the practical challenges behind the experience, from product matching and substitutions to store comparison, availability, mistakes like picking pasta when the user asked for rice, and the guardrails needed when a helpful assistant starts working with real baskets. About Yesid López: Yesid has a background spanning production ML, AI agents, generative AI, recommendation systems, and computer vision, he is passionate about turning intelligent ideas into reliable real-world applications. He is an avid hackathon participant and enjoys building GenAI applications in his spare time. Adversarial Review Loops in Practice at Doctolib by Dennis Loska, Software Engineer @ Dotolib Since agents can turn ideas into working features in minutes, developer work is shifting from writing to reviewing code. The main question is not whether it can be built, but whether it should be built. With an increase in code surface comes more bugs and more cognitive debt. Therefore setting up the proper development environment for agents and validating the output of those has become the focus! I'll showcase how my team utilizes techniques like adversarial review loops to create an agent friendly development environment or harness if you will. About Dennis Loska...