Stockholm MLOps Winterbash!!!

All right!!!

We're happy to announce that this year's Stockholm MLOps Winterbash!!! will happen December 11. This time, we'll be teaming up with IBM and once again, we'll round off the year with show-casing what some really amazing local start-ups are doing with AI.

***Please note that you will have to register at IBM's event page as well in order to attend (and be let in). Please do so here: ibm.biz/mlopswinterbash***

Also, AI Sweden will be onsite to record the presentations and publish them on their Youtube-channel! Once edited, they will appear here.

Event Program

- Doors open at 17:00 CET

- Talks begin at 17:45 CET

- There will be light food & drinks

Speaker Line-Up

• Stefan Palm, Solution Architect at IBM will give a talk titled "Build smarter AI with IBM Open source, from models to Generative Computing." Abstract: Let us explore the current status of IBM initiatives in the AI Open Source community. We will look at the most popular projects you should know about, understand where the latest Granite models may be an option, and also peek a bit into what might be a next way of using LLMs, using a new framework for what we call Generative Computing.
• Anders Hammarbäck, Co-Founder and Leonora Vesterbacka, Founding Data Scientist at Redpine AI will give a talk titled: "Unlocking the value of proprietary data for AI". Abstract: Redpine serves as the knowledge layer for agents. AI agents today rely on fragmented, low-quality, and often unlicensed public data — limiting accuracy, compliance, and real-world reliability. The core problem: agents need high-quality, rights-cleared, domain-specific knowledge, and that layer simply doesn’t exist. Redpine is solving this by providing a trusted knowledge layer where agents can access verified, licensed data on demand — enabling safer, more accurate, and compliant agentic AI.
• Henrik Skagerlind Fasth, Co-Founder of Lumoo Studio will give a talk titled: "Engineering Lumoo on Lovable: Full-Stack AI Beyond Expectations". Abstract: "How Lumoo was engineered on Lovable in ways that pushed the platform beyond expectations and powers everything from frontend, backend, Kubernetes, Docker, and integrations, enabling one of the most robust and flexible solutions in fashion, sport, and retail."
• Rasmus Sanne, CEO of Stardust will give a talk titled: "From hackathon project to €1.7M investment in 7 months". Abstract: Stardust is building the self-learning mind for marketing. AI that understands why ads sell by decoding the visual and verbal DNA behind high-performing creatives.
• Karoly Szipka, Founder & CEO of IPercept will give a talk titled "Unlocking the last generation of industrial machine efficiency". Abstract: Purpose-built for CNC (Computer Numerical Control) machines, IPercept boosts production and maintenance teams with an AI assistant that gives them the insights they need to improve machine reliability, quality, and capacity.
• Evgenii Kondratev, CPO at Talentium will give a talk titled: "Hiring in the AI Era: How Talentium is Rewriting the Hiring Playbook". Abstract: In a world where recruitment is broken—slow, biased, and built on outdated tools—AI has the potential to fundamentally reset how teams are built. In this session, Sebastian explores how Talentium’s AI agents transform hiring by combining human judgment with intelligent automation across sourcing, assessment, and candidate engagement. Participants will walk away with a clear understanding of how next-generation hiring works, why it restores the human in recruitment, and how modern teams can scale smarter, faster, and more sustainably with AI.
• Mika Eberstein Lundberg, AI Software Engineer at Sana will give a talk titled: "From Hype to Reality: Why We Built AI Slides Backwards (And What We Learned)". Abstract: While building AI-generated slide decks at Sana, our first approach went all-in on letting the model produce full HTML. The demos were beautiful, but in production it proved expensive, slow, insecure, and hard to control. This talk covers how we moved from impressive-but-chaotic HTML generation, to the structured JSON+HTML system we rely on today. We'll share what this journey taught us about building enterprise AI: why flashy flexibility doesn’t scale, how smart constraints boost reliability, and what it takes to ship AI features that teams can trust.

We look forward to seeing you there,

/Patrick & the Stockholm MLOps Team