Tricity AI Meetup #1
We planned a small first edition. Tricity apparently had other plans.
On May 20, Scalac.ai hosted the first Tricity AI Meetup in Gdańsk. The room was full, the talks were practical, and the conversations continued after the agenda ended.
Watch the talks, get the resource pack, and help us shape the next editions.
VIDEOS
Watch the talks
Three talks, three different angles on practical AI: local models, live agents, and security.
Local LLMs in practice
Piotr’s talk was not another generic overview of local LLMs.
It felt more like a „story of my life with local LLMs”, told through real use cases and actual engineering trade-offs.
Small spoiler: it started with a several hundred dollar API bill.
From there, Piotr walked through the messy parts of local inference: hardware choices, VRAM limits, inference software, serving stacks, privacy, cost, and the difference between „it runs” and „it makes sense”.
Google ADK and live AI agents
Mariusz did not just talk about AI agents.
He showed them live.
His session was a hands-on walkthrough of Google ADK, with agents being started, tested, and adjusted in real time.
He showed prompts, agent interactions, and what actually happens when separate agents start working together.
That made the talk concrete. Not „agents as an idea”, but agents as something you can run, inspect, and break a little.
AI agents in threat modeling
Paweł connected AI agents with threat modeling and security.
He showed where AI can help with application analysis, where it can mislead us, and why validation, evidence, and good context still matter.
A useful reality check from the talk: AI can be fast, but in security the answer still needs to be verified.
COMMUNITY
This meetup should be built with the community, not just for it.
Scalac.ai builds practical AI systems for companies that need more than a demo.
We work with AI agents, private LLMs, RAG systems, integrations, and production-ready AI workflows.
We organize Tricity AI Meetup because we want more local conversations about how AI is actually built, tested, secured, and deployed.