Compass Data Meetup by Lufthansa Systems Hungária

🚀 Join us at our first FREE Compass Data Meetup in partnership with Lufthansa Systems Hungária and Budapest Data Science Meetup! In the first talk, we'll explore how our speakers from Bitrise developed a caching simulation for their app development dashboard, overcoming challenges like event processing and historical data ordering to achieve efficient data management and real-time analysis. Then, the next presentation will showcase how Lufthansa System’s demand forecasting project evolved into a comprehensive MLOps ecosystem, offering AI deployment best practices and a blueprint for success across various industries.

📅Date & venue:

• 18 September 2024, Wednesday, from 6 pm
• Create26, Budapest, Király u. 26, 1061

To participate in the meetup you need to complete a short Google form!
🎟️ Register now for free to reserve your spot!🌐

Please note that the event's official language is English!

Schedule
5:30 - 6:00 Registration
6:00 - 6:40 First talk by Krisztina Nagy and Balázs Francsics
6:40 - 7:20 Second talk by Manuel Schmidt and Roland Madaras
7:20 - 9:00 Networking

Finding the balance of cost and performance in caching
by Krisztina Nagy (Bitrise, Data Analyst) and Balázs Francsics (Bitrise, Staff Data Engineer)
Caching in app development can significantly improve efficiency, but it introduces the challenge of balancing cost and performance. Our goal was to develop a dashboard that helps quickly identify this balance. We began by exploring different solutions, but encountered obstacles like processing a large volume of events and needing historical data in chronological order. In our final solution we implemented a caching simulation with infinite storage using dataflow and stream processing. This approach ensures efficient data management and real-time analysis for our dashboard. This presentation will cover our technical journey, the challenges we faced, and the solutions we implemented to optimize build caching.

From Prototype to Production: AI Best Practices Unveiled Through Demand Forecasting
by Manuel Schmidt (Lufthansa Systems) and Roland Madaras (Lufthansa Systems Hungária)
Bringing AI models to production is a complex journey. Our demand forecasting project serves as a case study to illustrate best practices and lessons learned in deploying AI at scale.
We aimed not just to improve predictions, but to establish a robust framework for operationalizing AI across our organization. Our solution evolved into a comprehensive MLOps ecosystem, incorporating cutting-edge practices in data engineering, model management, and performance monitoring.
This presentation will use demand forecasting as a lens to explore universally applicable AI best practices. We'll cover building scalable data pipelines, implementing version control for data and models, establishing KPIs, integrating DevOps principles, and ensuring model interpretability.

Join us to discover how the challenges of demand forecasting led us to develop a blueprint for AI success applicable across various domains and industries.

Sponsors and partners

• Create26: A coworking space in the heart of Budapest with a thriving community and extraordinary benefits.
• Lufthansa Systems Hungária
• Budapest Data Science Meetup