🛠️ Hands-On LLM Implementations and Predictive Workflows

On June 24, 2026, AI Tinkerers Karlsruhe is partnering with LLM Practitioners Karlsruhe and Blue Yonder for a technical build night focused on the practical application of foundation models. We are bypassing high-level slides and marketing pitches to focus on the engineering reality of building with LLMs, managing context, and deploying agentic workflows.

Hosted at Blue Yonder’s Karlsruhe office, this meetup brings together active developers, machine learning engineers, and researchers to share working code, system architectures, and the technical hurdles of orchestrating complex AI systems.

Our events are highly technical and selective. Space is strictly limited to 150 active builders to maintain a high-signal environment for peer-to-peer exchange.

📅 Schedule

🎙️ Submit Your Demo Proposal

We are looking for builders to showcase what they are actively working on. Whether you are optimizing local inference, building multi-agent orchestration pipelines, or tackling complex retrieval challenges, we want to see your code.

Your demo should answer: “How did you build this interesting thing?” rather than “Why should someone use this product?”

• Show, Don’t Tell: Run your project live. No slide decks are allowed (one architecture diagram is the maximum).
• Expose the Internals: Share your stack, prompt strategies, and the real failure modes you encountered.
• Teach, Don’t Sell: Focus on what another builder can reuse, adapt, or avoid.

Submit Your Demo Proposal

🥽 Who Should Attend

This event is curated exclusively for practitioners rolling up their sleeves to build with AI.

• ✅ Welcome: Software engineers, ML researchers, technical founders, and open-source contributors shipping code.
• 🚫 Not a fit: Recruiters, salespeople, marketers, consultants, or those looking for a general introduction to AI.

To ensure a high signal-to-noise ratio, all registrations are screened. Please provide links to your GitHub, LinkedIn, or technical portfolio when signing up.

📍 Venue & Logistics

The meetup is hosted by our sponsor, Blue Yonder, in Karlsruhe. To maintain a secure and focused environment, the exact entry details and access instructions will be sent automatically to approved attendees via their registration confirmation.

🤝 Sponsors & Partners

Blue Yonder is a leading provider of AI-driven supply chain platforms, unifying planning, execution, and commerce. Operating on a unified data cloud integrated with Snowflake, Blue Yonder utilizes a semantic knowledge graph and a modular agentic AI framework where specialized planning, inventory, and warehouse agents automate complex processes and resolve exceptions.

LLM Practitioners Karlsruhe is a local community dedicated to deep technical discussions, knowledge sharing, and collaboration around Large Language Models and generative AI technologies.

Interested in supporting the Karlsruhe builder ecosystem? View sponsorship opportunities.

📸 Experience AI Tinkerers

Event photos

AI Tinkerers - Karlsruhe

AI Tinkerers - Karlsruhe

AI Tinkerers - Karlsruhe

AI Tinkerers - Karlsruhe

AI Tinkerers - Karlsruhe

AI Tinkerers - Karlsruhe

AI Tinkerers Karlsruhe Stats

• Attendees: This community of 31 elite professionals comprises 65% LLM/RAG specialists and 58% machine learning engineers. With 35% holding executive leadership roles, the group bridges high-level strategy with technical execution. Notable for its deep integration with KIT research and industrial giants like SAP and Porsche, the membership excels in transitioning complex scientific R&D into production-ready agentic AI systems.
• Companies Represented: Innovative AI and automotive leaders like Porsche Engineering, Cinemo, and Netze BW, alongside high-growth startups such as Renumics, bloop, i22, Exxeta, Recall Space, Pryvet.ai, datin, and more.
• Demos: 1 demo has been submitted and 1 has been presented. The most exciting themes have focused on agentic AI for engineering data analysis, practical data management with DuckDB, and tool-based architectures paired with code generation. Stefan Suwelack’s “Agentic AI for Engineering Data Analysis” has especially highlighted how to streamline sensor data handling and automate industrial analytics end-to-end.
• Testimonials:

  “I really appreciated the quality of the discussions at AI Tinkerers - Karlsruhe… It’s a place where you can ask questions, share what you’ve built, and get thoughtful feedback.”

  Mehmet A., Stuttgart — 2025-07-31

  “AI Tinkerers - Karlsruhe struck the right balance between technical depth and approachability… I enjoyed the people and the community vibe as much as the content.”

  Elena R., Karlsruhe — 2025-06-22

  “AI Tinkerers - Karlsruhe was a fantastic meetup… The conversations were insightful, hands‑on, and genuinely encouraging. I left with new ideas and motivation.”

  Sara K., Karlsruhe — 2025-09-28

  “I enjoyed the friendly and very practical atmosphere… It was a great way to meet like‑minded people and learn from them.”

  Alex M., Karlsruhe — 2025-10-12

  “The meetup has a welcoming community and a strong focus on practical AI topics… The group feels very active and collaborative, not just theoretical.”

  David L., Heidelberg — 2025-08-19

  “This meetup consistently delivers value… It’s easy to join, the organizers create a great atmosphere, and the attendees are both curious and supportive.”

  Jonas P., Karlsruhe — 2025-05-14

A great demo, consistent with the speaker-form guidance and the characteristics of the highest-rated example, is one that demonstrates real implementation details in a short window: show how the system works internally (architecture and data flow), emphasize specific technical discoveries (“why this storage/query stack,” “how context is shaped,” “what changed in the agent workflow”), and include at least one concrete mechanism the audience can understand (e.g., code generation like txt2sql driving actual queries, or an AI-first UI that exposes the system’s reasoning/actions). To avoid common reasons demos underperform, don’t rely on marketing-style language, avoid purely conceptual explanations, and don’t hide the engineering “how” behind generic descriptions. Even when the rating data provided is limited, the single strong example indicates that audience members reward demos that expose the stack, tradeoffs, and practical design choices—especially those that translate AI behavior into something domain-relevant and inspectable.

The demo “Agentic AI for Engineering Data Analysis” by Stefan Suwelack in Karlsruhe is presented as a hands-on walkthrough of Onion, an agentic AI system for mechanical engineers analyzing test-, fleet-, and production data; it specifically focuses on how sensor data is stored and queried (Parquet + DuckDB), how context is shaped in the AI harness, and how the architecture evolved from a tool-based approach to a hybrid setup that includes code generation (txt2sql). This is likely why it earned a 4/5 rating: the talk’s topics are grounded in concrete engineering mechanisms and system components (data infrastructure, AI orchestration, and AI-first UI/UX), which makes the session feel like a genuine technical reveal instead of a high-level concept pitch. See the demo page here: Agentic AI for Engineering Data Analysis.