Chips and compute options for DACH: Cerebras, SiPearl/EU landscape, GPUs, cloud vs on-prem, TCO and energy.

In this episode, I sat down with tech humanist Kate O’Neill to explore how organizations can balance human-centered design in a time when everyone is racing to find ways to leverage AI in their businesses. Kate introduced her “Now–Next Continuum,” a framework that distinguishes digital transformation (catching up) from true innovation (looking ahead). We dug into real-world challenges and tensions of moving fast vs. creating impact with AI, how ethics fits into decision making, and the role of data in making informed decisions. 

Kate stressed the importance of organizations having clear purpose statements and values from the outset, proxy metrics she uses to gauge human-friendliness, and applying a “harms of action vs. harms of inaction” lens for ethical decisions. Her key point: human-centered approaches to AI and technology creation aren’t slow; they create intentional structures that speed up smart choices while avoiding costly missteps.

Highlights/ Skip to:

How Kate approaches discussions with executives about moving fast, but also moving in a human-centered way when building out AI solutions (1:03) Exploring the lack of technical backgrounds among many CEOs and how this shapes the way organizations make big decisions around technical solutions (3:58)  FOMO and the “Solution in Search of a Problem” problem in Data (5:18)  Why ongoing ethnographic research and direct exposure to users are essential for true innovation (11:21)  Balancing organizational purpose and human-centered tech decisions, and why a defined purpose must precede these decisions (18:09) How organizations can define, measure, operationalize, and act on ethical considerations in AI and data products (35:57) Risk management vs. strategic optimism: balancing risk reduction with embracing the art of the possible when building AI solutions (43:54)

Quotes from Today’s Episode "I think the ethics and the governance and all those kinds of discussions [about the implications of digital transformation] are all very big word - kind of jargon-y kinds of discussions - that are easy to think aren't important, but what they all tend to come down to is that alignment between what the business is trying to do and what the person on the other side of the business is trying to do." –Kate O’Neill

" I've often heard the term digital transformation used almost interchangeably with the term innovation. And I think that that's a grave disservice that we do to those two concepts because they're very different. Digital transformation, to me, seems as if it sits much more comfortably on the earlier side of the Now-Next Continuum. So, it's about moving the past to the present… Innovation is about standing in the present and looking to the future and thinking about the art of the possible, like you said. What could we do? What could we extract from this unstructured data (this mess of stuff that’s something new and different) that could actually move us into green space, into territory that no one’s doing yet? And those are two very different sets of questions. And in most organizations, they need to be happening simultaneously." –Kate O’Neill

"The reason I chose human-friendly [as a term] over human-centered partly because I wanted to be very honest about the goal and not fall back into, you know, jargony kinds of language that, you know, you and I and the folks listening probably all understand in a certain way, but the CEOs and the folks that I'm necessarily trying to get reading this book and make their decisions in a different way based on it." –Kate O’Neill

“We love coming up with new names for different things. Like whether something is “cloud,” or whether it’s like, you know, “SaaS,” or all these different terms that we’ve come up with over the years… After spending so long working in tech, it is kind of fun to laugh at it. But it’s nice that there’s a real earnestness [to it]. That’s sort of evergreen [laugh]. People are always trying to genuinely solve human problems, which is what I try to tap into these days, with the work that I do, is really trying to help businesses—business leaders, mostly, but a lot of those are non-tech leaders, and I think that’s where this really sticks is that you get a lot of people who have ascended into CEO or other C-suite roles who don’t come from a technology background.” 

–Kate O’Neill

"My feeling is that if you're not regularly doing ethnographic research and having a lot of exposure time directly to customers, you’re doomed. The people—the makers—have to be exposed to the users and stakeholders.  There has to be ongoing work in this space; it can't just be about defining project requirements and then disappearing. However, I don't see a lot of data teams and AI teams that have non-technical research going on where they're regularly spending time with end users or customers such that they could even imagine what the art of the possible could be.”

–Brian T. O’Neill

Links

KO Insights: https://www.koinsights.com/ LinkedIn for Kate O’Neill: https://www.linkedin.com/in/kateoneill/ Kate O’Neill Book: What Matters Next: A Leader's Guide to Making Human-Friendly Tech Decisions in a World That's Moving Too Fast

Overview of Google I/O Extended 2025 with live-streamed sessions and hands-on workshops for Android, Web, Cloud, and more.

At Berlin Buzzwords, industry voices highlighted how search is evolving with AI and LLMs.

• Kacper Łukawski (Qdrant) stressed hybrid search (semantic + keyword) as core for RAG systems and promoted efficient embedding models for smaller-scale use.
• Manish Gill (ClickHouse) discussed auto-scaling OLAP databases on Kubernetes, combining infrastructure and database knowledge.
• André Charton (Kleinanzeigen) reflected on scaling search for millions of classifieds, moving from Solr/Elasticsearch toward vector search, while returning to a hands-on technical role.
• Filip Makraduli (Superlinked) introduced a vector-first framework that fuses multiple encoders into one representation for nuanced e-commerce and recommendation search.
• Brian Goldin (Voyager Search) emphasized spatial context in retrieval, combining geospatial data with AI enrichment to add the “where” to search.
• Atita Arora (Voyager Search) highlighted geospatial AI models, the renewed importance of retrieval in RAG, and the cautious but promising rise of AI agents.

Together, their perspectives show a common thread: search is regaining center stage in AI—scaling, hybridization, multimodality, and domain-specific enrichment are shaping the next generation of retrieval systems.

Kacper Łukawski Senior Developer Advocate at Qdrant, he educates users on vector and hybrid search. He highlighted Qdrant’s support for dense and sparse vectors, the role of search with LLMs, and his interest in cost-effective models like static embeddings for smaller companies and edge apps. Connect: https://www.linkedin.com/in/kacperlukawski/

Manish Gill
Engineering Manager at ClickHouse, he spoke about running ClickHouse on Kubernetes, tackling auto-scaling and stateful sets. His team focuses on making ClickHouse scale automatically in the cloud. He credited its speed to careful engineering and reflected on the shift from IC to manager.
Connect: https://www.linkedin.com/in/manishgill/

André Charton
Head of Search at Kleinanzeigen, he discussed shaping the company’s search tech—moving from Solr to Elasticsearch and now vector search with Vespa. Kleinanzeigen handles 60M items, 1M new listings daily, and 50k requests/sec. André explained his career shift back to hands-on engineering.
Connect: https://www.linkedin.com/in/andrecharton/

Filip Makraduli
Founding ML DevRel engineer at Superlinked, an open-source framework for AI search and recommendations. Its vector-first approach fuses multiple encoders (text, images, structured fields) into composite vectors for single-shot retrieval. His Berlin Buzzwords demo showed e-commerce search with natural-language queries and filters.
Connect: https://www.linkedin.com/in/filipmakraduli/

Brian Goldin
Founder and CEO of Voyager Search, which began with geospatial search and expanded into documents and metadata enrichment. Voyager indexes spatial data and enriches pipelines with NLP, OCR, and AI models to detect entities like oil spills or windmills. He stressed adding spatial context (“the where”) as critical for search and highlighted Voyager’s 12 years of enterprise experience.
Connect: https://www.linkedin.com/in/brian-goldin-04170a1/

Atita Arora
Director of AI at Voyager Search, with nearly 20 years in retrieval systems, now focused on geospatial AI for Earth observation data. At Berlin Buzzwords she hosted sessions, attended talks on Lucene, GPUs, and Solr, and emphasized retrieval quality in RAG systems. She is cautiously optimistic about AI agents and values the event as both learning hub and professional reunion.
Connect: https://www.linkedin.com/in/atitaarora/

Overview of Le Wagon’s Data Engineering bootcamp, including program structure, flexible learning formats, and financing options. Topics include data infrastructure, cloud services, and data pipelines; career support during and after the bootcamp; and guidance on choosing the right learning format.

EBMT, one of the biggest medical registries in Europe, has rebuilt its core data system from scratch, after 20 years of service, to keep up with growing data needs, modern technologies, and the evolving needs of researchers in blood and marrow transplantation. The new AWS-based system supports data collection and analysis at scale, using cloud infrastructure and business intelligence tools to improve data quality and data usability across EBMT’s network.

Air France KLM partnered with Datashift to tackle data silos, regulatory complexity, and fragmented access by building a central data marketplace. Using Collibra and Google Cloud, they now empower the organization with trusted, governed, and self-service data to support critical operations and decision-making.

Sligro Food Group, Dutch market leader in food service, needed to centralise data to improve retail decision-making and stay competitive. Moving away from on-premise databases, they simplified integration to GCP. Join this session to learn why centralised data is key for Sligro, why enterprises choose Fivetran over legacy tools, and how to integrate data into the cloud for real-time analytics.

Veel organisaties worstelen met een versnipperd datalandschap vol scripts en ETL-tooling die alleen door experts begrepen wordt. Ploeger Logistics laat zien dat het anders kan. Samen met Infotopics migreerde de logistiek dienstverlener haar complete data-logistiek naar de cloud – zonder verlies van continuïteit. Het resultaat: één schaalbaar platform, met een beter datafundament voor de hele organisatie. Tijdens deze sessie ontdek je de keuzes, obstakels en impact van deze transformatie.

Discover how to design and deploy powerful multi-agent systems on Google Cloud. Cloud & AI Consultant Timothy van der Werf demonstrates how agents collaborate, share tasks, and autonomously handle complex processes using technologies like Agentspace. This session delivers practical examples and insights for anyone looking to apply AI to data-driven innovation, customer engagement, or operational optimization.

Unlock efficient AI! This playbook explores Small Language Models (SLMs) as a cost-effective alternative to large LLMs. Learn to select, deploy (local/cloud), and utilize these powerful, often open-source models for high-value, targeted tasks.

Elliot Foreman and Andrew DeLave from ProsperOps joined Yuliia and Dumky to discuss automated cloud cost optimization through commitment management. As Google go-to-market director and senior FinOps specialist, they explain how their platform manages over $4 billion in cloud spend by automating reserved instances, committed use discounts, and savings plans across AWS, Azure, and Google Cloud. The conversation covers the psychology behind commitment hesitation, break-even point mathematics for cloud discounts, workload volatility optimization, and why they avoid AI in favor of deterministic algorithms for financial decisions. They share insights on managing complex multi-cloud environments, the human vs automation debate in FinOps, and practical strategies for reducing cloud costs while mitigating commitment risks.

The official guide to MongoDB architecture, tools, and cloud features, written by leading MongoDB subject matter experts to help you build secure, scalable, high-performance applications Key Features Design resilient, secure solutions with high performance and scalability Streamline development with modern tooling, indexing, and AI-powered workflows Deploy and optimize in the cloud using advanced MongoDB Atlas features Purchase of the print or Kindle book includes a free PDF eBook Book Description Delivering secure, scalable, and high-performance applications is never easy, especially when systems must handle growth, protect sensitive data, and perform reliably under pressure. The Official MongoDB Guide addresses these challenges with guidance from MongoDB’s top subject matter experts, so you learn proven best practices directly from those who know the technology inside out. This book takes you from core concepts and architecture through to advanced techniques for data modeling, indexing, and query optimization, supported by real-world patterns that improve performance and resilience. It offers practical coverage of developer tooling, IDE integrations, and AI-assisted workflows that will help you work faster and more effectively. Security-focused chapters walk you through authentication, authorization, encryption, and compliance, while chapters dedicated to MongoDB Atlas showcase its robust security features and demonstrate how to deploy, scale, and leverage platform-native capabilities such as Atlas Search and Atlas Vector Search. By the end of this book, you’ll be able to design, build, and manage MongoDB applications with the confidence that comes from learning directly from the experts shaping the technology. What you will learn Build secure, scalable, and high-performance applications Design efficient data models and indexes for real workloads Write powerful queries to sort, filter, and project data Protect applications with authentication and encryption Accelerate coding with AI-powered and IDE-based tools Launch, scale, and manage MongoDB Atlas with confidence Unlock advanced features like Atlas Search and Atlas Vector Search Apply proven techniques from MongoDB's own engineering leaders Who this book is for This book is for developers, database professionals, architects, and platform teams who want to get the most out of MongoDB. Whether you’re building web apps, APIs, mobile services, or backend systems, the concepts covered here will help you structure data, improve performance, and deliver value to your users. No prior experience with MongoDB is required, but familiarity with databases and programming will be helpful.

Many industrial and legacy systems still push critical data over TCP streams. Instead of reaching for heavyweight cloud platforms, you can build fast, lean batch pipelines on-prem using Python and DuckDB.

In this talk, you'll learn how to turn raw TCP streams into structured data sets, ready for analysis, all running on-premise. We'll cover key patterns for batch processing, practical architecture examples, and real-world lessons from industrial projects.

If you work with sensor data, logs, or telemetry, and you value simplicity, speed, and control this talk is for you.