From Foundation Models to Real-World AI & The Art & Science of Talking to AI​

From Foundation Models to Real-World AI & The Art & Science of Talking to AI

Join us for our next NLP Zurich tech shindig at Zalando! For this event, you will also need to register here to ensure you can enter the Zalando office. Please register by 2nd March.
We are looking forward to welcoming you!

Agenda
17:45 Join the meetup at Zalando
18:00 Michael Gygli (Zalando): From Foundation Models to Real-World AI: Lessons from E-Commerce Fashion Discovery
18:30 Kemeng Zheng (AWS): The Art & Science of Talking to AI: Prompt Engineering Best Practices for LLMs
20:00 Networking ⊂(◉‿◉)つ

Location: Zalando Switzerland AG Prime Tower Hardstrasse 201 8005 Zurich
Reminder: You need to register by 2nd March to be able to enter the event building: https://forms.gle/viYCDsTuN1Bgj2dy7

From Foundation Models to Real-World AI: Lessons from E-Commerce Fashion Discovery

Talk summary
Recent advancements in foundation models like CLIP, ChatGPT, and DeepSeek have unlocked a new level of capabilities, reaching a human-level understanding in a wide range of tasks. But transforming these raw models into high-performing, real-world solutions takes more than just API calls.
This talk explores how to use CLIP, ChatGPT and the like, for specific use cases, exemplified on the use case of fashion discovery. We’ll start with diving into strategies for curating and generating the right training and evaluation data - whether through interaction signals, human feedback, or LLMs as judges. And we’ll address the MLOps challenges such as deployment, measuring quality in production, as well as iteratively making improvements.
By the end of this talk, you'll have a well-defined roadmap for going from tinkering with foundation models to using them for impactful, real-world advancements.

About the Speaker
Michael Gygli is a Principal Applied Scientist at Zalando. He has a track record in shipping ML-enabled products and pushing ML research forward. He started his career with a Ph.D. in Computer Vision from ETH Zurich, gained industry experience at Google and Yahoo and has co-founded Cerrion. At Cerrion, a YCombinator startup, he built the initial MVP and served as the CTO for 2.5 years, leading a team of six engineers in backend development and machine learning, before transitioning from an operational to an advisory role.
As a researcher, he has published 30+ papers with 3000+ citations. His research interests lay in textual-visual learning and foundation models and self-improving models.

The Art & Science of Talking to AI: Prompt Engineering Best Practices for LLMs

Talk summary
Large Language Models are transforming how we work—but their effectiveness depends heavily on how we communicate with them. This session explores both the creative art and technical science of prompt engineering, showing you how to craft prompts that deliver better results with AWS AI services and other LLMs. Learn practical techniques for improving accuracy, maintaining context, and optimizing responses, whether you're building enterprise AI solutions or enhancing daily AI interactions.

About the Speaker
Hi, I’m Kemeng Zhang, a Solutions Architect at AWS. With a background in AI/ML and cloud architecture, I help startups and enterprises design scalable, cost-efficient solutions that turn ideas into reality. Previously, I worked as an AI Application Architect, building machine learning-driven applications across industries like finance, healthcare, and manufacturing.

About our Host
Zalando is one of Europe’s leading online platforms for fashion and lifestyle, offering a seamless connection between customers, brands, and partners across 25 markets. It serves over 50 million active customers with a wide variety of fashion and lifestyle products, including clothing, shoes, accessories, and beauty items. Zalando's approach integrates advanced digital solutions to enhance various aspects of the fashion industry. This allows them to deliver a personalized and innovative shopping experience.
Teams at Zalando Zürich Tech Hub work on solving some of the most important challenges in the fashion industry, such as finding the right size or helping customers navigate Zalando’s comprehensive assortment in order to find what they love. Through a full stack of professionals in software development, computer vision, applied scientist, researchers and product managers. Zalando is turning visionary ideas into new emerging businesses with real potential. By addressing complex challenges within the fashion sector, Zalando aims to revolutionize the digital shopping experience and drive forward its goal of creating the most inspiring and personalized fashion platform for customers.

NLP Zurich:
Meetup.com: https://www.meetup.com/NLP-Zurich/
Linkedin: https://www.linkedin.com/company/nlp-zurich