As always, there will be plenty of time for networking, discussions, and snacks & drinks courtesy of our host.
đź“Ť Event Details
đź“… Date: Thursday, May 28th, 2026
📍 Venue: Giesinger Bräu Werk2 sponsored by aiven
đź“« Address: DetmoldstraĂźe 40, 80935 MĂĽnchen
🎤 PostgreSQL, ClickHouse, or OpenSearch? A Practical Guide to Picking a Vector Strategy for Production AI
by Geethu Uday, Aiven
There is no 'best' vector database, only the right one for your specific architecture. Choosing a database for AI features usually starts simple, but gets complicated fast when you move toward production. The common struggle is figuring out whether to stick with a database you already know or add a specialized vector store that increases your operational overhead.
This session looks at the engineering trade-offs between Postgres, ClickHouse, and OpenSearch when handling vector data. We’ll compare their indexing strategies, memory footprints, and how they handle the 'Metadata Filtering' challenge at scale. The goal is to provide a straightforward framework for picking a strategy based on your data scale, query patterns, and how much time you actually want to spend on maintenance.
About Geethu Uday:
Geethu Uday is a Solutions Architect at Aiven, specializing in open-source data infrastructure and multi-cloud architecture. With a background in enterprise data engineering and an MSc in Data Science, she helps teams design scalable streaming and analytical platforms using PostgreSQL, ClickHouse, OpenSearch and Kafka. She partners with engineering and platform teams to help them make smarter decisions around scalability, cost, and architecture and is currently exploring the practical implementation of AI within the data stack.
🎤 From Sound to Satellite: How We Combine Bioacoustics and Remote Sensing for Biodiversity Insights
by David Schmider, Hula Earth & Konrad Heidler, Hula Earth
Measuring biodiversity is surprisingly difficult: data is often fragmented, non-standardized, and rarely available in real time. We show how at Hula Earth we combine bioacoustics (e.g., bird calls) with satellite data to generate scalable insights about ecosystems.
We start with the problem and the perspective behind it: why this matters in the first place and where current approaches hit their limits. Then we give an overview of how our acoustic edge systems work and how the data flows into pipelines and ML models. In the second part, we go a bit deeper — from signal processing and feature extraction to methods for fusion with geodata and integration into a near-production system.
About David Schmider:
David is Co-Founder and CTO of Hula Earth. He comes from a software/DevOps background and has worked across pretty much every layer of the stack. At Hula Earth, he's responsible for bringing the entire tech side together — from hardware and firmware to data science and software — and building it all into a functioning end-to-end system. Background: B.Sc. in Computer Science.
About Konrad Heidler:
Konrad is Senior Data Scientist at Hula Earth. He has a background in mathematics/data science and a PhD in geodata processing. Previously, he led a research group at TUM. At Hula Earth, he focuses on everything related to models, data, and pipelines — from finding data sources to modeling and integrating them into our systems, in both bioacoustics and remote sensing.
đź“‹ Agenda
18:30 - Doors open & networking
18:40 - Welcome & introductions
18:45 - Talk #1: PostgreSQL, ClickHouse, or OpenSearch? A Practical Guide to Picking a Vector Strategy for Production AI
19:30 - Pizza & networking
20:00 - Talk #2: From Sound to Satellite: How We Combine Bioacoustics and Remote Sensing for Biodiversity Insights
20:45 - Open networking
